DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment
This action is responsive to the request for continued examination (RCE), amendments and remarks received 01 April 2026. Claims 1 - 3, 5 - 13, 15 - 17, 19 and 20 are currently pending.
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 01 April 2026 has been entered.
Claim Objections
Claim 5 is objected to because of the following informalities: Lines 13 - 14 of claim 5 recite, in part, “each of the one or more known products” which appears to contain inconsistent claim terminology and/or a minor informality. The Examiner suggests amending the claim to --each of themaintain consistency with line 9 of claim 5 and to improve the clarity and precision of the claim. Appropriate correction is required.
Claim 15 is objected to because of the following informalities: Lines 8 - 9 of claim 15 recite, in part, “perform a comparison, by the machine learning model, at least digital pixel data” which appears to contain a grammatical error and/or a minor informality. The Examiner suggests amending the claim to --perform a comparison, by the machine learning model, of at least digital pixel data-- in order to improve the clarity and precision of the claim. Appropriate correction is required.
Claim 15 is objected to because of the following informalities: Lines 11 - 14 of claim 15 recite, in part, “generate, by the machine learning model, for each of one or more known product images in the group of known product images, a similarity score between the target product image and one or more known product images from the group of known product images” which appears to contain a grammatical error and/or a minor informality. The Examiner suggests amending the claim to --generate, by the machine learning model, for each of one or more known product images in the group of known product images, a similarity score between the target product image and the known product image-- in order to improve the clarity and precision of the claim. Appropriate correction is required.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1 - 3, 5 - 13, 15 - 17, 19 and 20 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 1 recites the limitation "the machine learning model executed by the one or more processors" (emphasis added) in lines 5 - 6. There is insufficient antecedent basis for this limitation in the claim.
Claim 1 recites the limitation "the one or more processors" in lines 5 - 6. There is insufficient antecedent basis for this limitation in the claim.
Claim 1 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention because it is unclear as to which group of known object images “the group of known object images” recited on line 17 is referencing. Is it referring to the “group of known object images” recited on line 9 of claim 1 or the “group of known object images” recited on lines 11 - 12 of claim 1? Additionally, it is unclear as to whether the “group of known object images” recited on line 9 of claim 1 and the “group of known object images” recited on lines 11 - 12 of claim 1 are the same or different. Clarification and appropriate correction are required. For purposes of examination the Examiner will treat “the group of known object images” recited on line 17 of claim 1 as referencing the “group of known object images” recited on lines 11 - 12 of claim 1.
Claim 1 recites the limitation "the digital pixel data of the given known object image" in lines 25 - 26. There is insufficient antecedent basis for this limitation in the claim.
Claim 1 recites the limitation "the known object images that are more similar to the target object" in lines 27 - 28. There is insufficient antecedent basis for this limitation in the claim.
Claim 1 recites the limitation "the one or more known object images having a similarity score above a threshold similarity score" (emphasis added) in lines 31 - 32. There is insufficient antecedent basis for this limitation in the claim.
Claim 1 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention because it is unclear as to which similar object image “the respective similar object image” recited on line 42 is referencing since line 30, lines 34 - 36 and lines 37 - 38 of claim 1 make it clear that there are a set of similar object images and that there is historical event data associated with each similar object image in the set of similar object images. Therefore, the Examiner asserts that it is unclear as to which similar object image in the set of similar object images “the respective similar object image” recited on line 42 of claim 1 is referencing. Clarification and appropriate correction are required. For purposes of examination the Examiner will treat “the respective similar object image” recited on line 42 of claim 1 as referencing any similar object image in the set of similar object images.
Claim 5 recites the limitation "the similarity scores generated for each of the one or more known products" in lines 12 - 13. There is insufficient antecedent basis for this limitation in the claim.
Claim 8 recites the limitation "the known product images that are more similar to the target product image" in lines 12 - 13. There is insufficient antecedent basis for this limitation in the claim.
Claim 8 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention because it is unclear as to which similar product image “the respective similar product image” recited on line 23 is referencing since line 7, lines 15 - 17 and lines 18 - 21 of claim 8 make it clear that there are a set of similar product images and that there is historical event data associated with each similar product image in the set of similar product images. Therefore, the Examiner asserts that it is unclear as to which similar product image in the set of similar product images “the respective similar product image” recited on line 23 of claim 8 is referencing. Clarification and appropriate correction are required. For purposes of examination the Examiner will treat “the respective similar product image” recited on line 23 of claim 8 as referencing any similar object image in the set of similar product images.
Claim 8 recites the limitation "the predicted characteristic for the target product" in lines 26 - 27. There is insufficient antecedent basis for this limitation in the claim.
Claim 12 recites the limitation "the similarity scores between the target product (i) and each of the known products" in lines 12 - 13. There is insufficient antecedent basis for this limitation in the claim.
Claim 15 recites the limitation "the known product images that are more similar to the target product image" in lines 16 - 17. There is insufficient antecedent basis for this limitation in the claim.
Claim 15 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention because it is unclear as to which similar product image “the respective similar product image” recited on line 30 is referencing since line 19, lines 22 - 24 and lines 25 - 28 of claim 15 make it clear that there are a set of similar product images and that there is historical event data associated with each similar product image in the set of similar product images. Therefore, the Examiner asserts that it is unclear as to which similar product image in the set of similar product images “the respective similar product image” recited on line 30 of claim 15 is referencing. Clarification and appropriate correction are required. For purposes of examination the Examiner will treat “the respective similar product image” recited on line 30 of claim 15 as referencing any similar object image in the set of similar product images.
Claim 17 recites the limitation "the predicted characteristic for the target product" in lines 4 - 5. There is insufficient antecedent basis for this limitation in the claim.
Claim 20 recites the limitation "the predicted characteristic" in line 4. There is insufficient antecedent basis for this limitation in the claim.
Claims 2, 3, 6, 7, 9 - 11, 13, 16 and 19 are also rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, due to being dependent upon a rejected base claim but would be withdrawn from the rejection if their base claim overcomes the rejection.
Claim Rejections - 35 USC § 101
The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
The rejections to claims 1 - 6 and 8 - 20 under 35 U.S.C. 101 are hereby withdrawn in view of the amendments and remarks received 01 April 2026.
Response to Arguments
Applicant’s arguments with respect to claim(s) 1 - 3, 5 - 13, 15 - 17, 19 and 20 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Applicant's arguments filed 01 April 2026 have been fully considered but they are not persuasive.
On pages 21 - 23 of the remarks the Applicant’s Representative argues that the combination of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. fails to teach or suggest “weighting, for each similar object image in the set of similar object images, the historical event data associated with the respective similar object image with the similarity score associated with the respective similar object image". The Applicant’s Representative argues that since “the Examiner uses Craparotta's ‘distance between features’ as equating to a similarity score” that “the references must teach or suggest Craparotta's ‘distance’ being used to weight historical event data of a respective similar object image.” First, the Applicant’s Representative argues that “there is no correlation between Craparotta's ‘distance’ between features” and the weighted aggregation on time-series of Ekambaram et al. and that there “is no teaching or suggestion to one of ordinary skill in the art to combine them in any way.” Second, the Applicant’s Representative argues that using the “distance” from Craparotta et al. “as a weighting function would cause more similar feature vectors to receive less weight and be less relevant.” Additionally, the Applicant’s Representative argues that Ekambaram et al. fail to cure the deficiencies of Craparotta et al. at least because there “is no teaching or suggestion to perform weighting by ‘similarity score’” and because “[p]erforming weighted aggregation on time-series of k nearest neighbors’ fails to teach or suggest ‘weighting... the historical event data associated with the respective similar object image with the similarity score associated with the respective similar object image’ as recited in claim 1.” Therefore, the Applicant’s Representative argues that the previously cited prior art references, individually or in combination, fail to teach or suggest the aforementioned disputed claim limitation(s).
The Examiner respectfully disagrees.
The Examiner asserts that, at least, Ekambaram et al. disclose utilizing a distance between features to perform their weighted aggregation on time-series, see at least pages 3112 - 3113 section 3.1 and page 3116 section 4.2 paragraph 1 of Ekambaram et al. wherein they disclose fetching “k-nearest neighbors from the historical data”, that “estimator
ℇ
a produces a sales time-series estimate by performing weighted aggregation on time-series of k nearest neighbors”, that Embedding KNN can represent “the product features as a vector embedding. The vector embedding defines a proper distance metric. It is expected that similar products are close to each other compared to dissimilar products in this embedding space”, that “X𝑝, and
Y
𝑝 represents the image/unstructured data and sales time-series attributed to product p. 𝑑 (Φ(X𝑝𝑖), Φ(X𝑝𝑗)) is the distance metric between product p𝑖, and p𝑗 in the embedding space”, that “similar to attribute-based KNN, a KNN estimator ƐI is defined as
PNG
media_image1.png
60
420
media_image1.png
Greyscale
” and that “Embedding KNN defines the similarity of products based on the cosine distance between the product image embeddings”. Thus, the Examiner asserts that there is a correlation between Craparotta's ‘distance’ between features” and the weighted aggregation on time-series of Ekambaram et al. and that at least Ekambaram et al. provide a teaching and suggestion to utilize a distance between features to weight historical event data of a respective similar object image.
Additionally, the Examiner asserts that using “distance” as a weighting function in the estimator ƐI function defined in section 3.1.2 of Ekambaram et al. would not cause more similar feature vectors to receive less weight at least because the denominator in the estimator ƐI function defined in section 3.1.2 of Ekambaram et al. equals the summation of each distance of all the nearest neighboring vectors up to the distance of a current nearest neighboring vector, and thus grows as more nearest neighboring vectors are considered, whereas the numerator only ever equals the distance of a current nearest neighboring vector being considered. Thus, the Examiner asserts that more similar products with smaller distances receive more weight in the estimator ƐI function defined in section 3.1.2 of Ekambaram et al.
Lastly, the Examiner asserts that, at least, Ekambaram et al. disclose and suggest “weighting…historical event data associated with the respective similar object image with the similarity score associated with the respective similar object image” at least because, as shown herein above and in the cited portions, Ekambaram et al. disclose that “Embedding KNN defines the similarity of products based on the cosine distance between the product image embeddings”, that it “is expected that similar products are close to each other compared to dissimilar products in this embedding space” and because the estimator ƐI function defined in section 3.1.2 of Ekambaram et al. weights the sales time-series attributed to the nearest k neighboring products with their respective distance metrics. The Examiner asserts that the distance metric of Ekambaram et al. corresponds to a similarity score at least because Ekambaram et al. disclose that it is utilized to define the similarity between products.
Therefore, the Examiner asserts that the previously cited prior art references teach and suggest the aforementioned disputed claim limitation(s).
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
Claims 1 - 3, 5, 8 - 10, 12 and 14 - 17 are rejected under 35 U.S.C. 103 as being unpatentable over Giuseppe Craparotta, Sébastien Thomassey, Amedeo Biolatti, "A Siamese Neural Network Application for Sales Forecasting of New Fashion Products Using Heterogeneous Data", International Journal of Computational Intelligence Systems, Vol. 12(2), Nov. 2019, pages 1537 - 1546, herein referred to as “Craparotta et al.”, in view of Tibau-Puig et al. U.S. Publication No. 2019/0294879 A1 in view of Vijay Ekambaram, Kushagra Manglik, Sumanta Mukherjee, Surya Shravan Kumar Sajja, Satyam Dwivedi, Vikas Raykar, "Attention based Multi-Modal New Product Sales Time-series Forecasting", Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, Aug. 2020, pages 3110 - 3118, herein referred to as “Ekambaram et al.”, in view of Ikeda et al. U.S. Publication No. 2005/0238209 A1.
- With regards to claim 1, Craparotta et al. disclose a method, (Craparotta et al., Pg. 1537 Abstract, Pg. 1540 § 3 ¶ 1, Pg. 1540 Fig. 5, Pgs. 1544 - 1545 § 5) the method comprising: inputting, to a machine learning model, a target object image in digital form that represents a target object; (Craparotta et al., Pg. 1538 Fig. 2, Pg. 1539 § 2.1 - § 2.2, Pg. 1539 Figs. 3 & 4, Pg. 1540 Fig. 5, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1541 Fig. 7, Pg. 1542 § 4.1 ¶ 1 - 2, Pg. 1543 Fig. 11, Pg. 1544 § 5, Pg. 1545 Fig. 16) extracting, by the machine learning model, digital pixel data of the target object image; (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph, Pg. 1539 § 2.1, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1541 Fig. 7) processing, by the machine learning model, the digital pixel data of the target object image by pixel analysis and comparing to pixel arrays of digital pixel data of a group of known object images; (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph - Pg. 1539 § 2.2, Pg. 1538 Fig. 2, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1541 Fig. 7, Pg. 1542 § 4 - § 4.1 ¶ 2, Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pg. 1544 § 5 ¶ 1, Pg. 1545 Fig. 16) performing a comparison, by the machine learning model, of the digital pixel data of the target object image to digital pixel data associated with a group of known object images; (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph - Pg. 1539 § 2.2, Pg. 1538 Fig. 2, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1541 Fig. 7, Pg. 1542 § 4 - § 4.1 ¶ 2, Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pg. 1544 § 5 ¶ 1, Pg. 1545 Fig. 16) generating, by the machine learning model, for each of one or more known object images in the group of known object images, a respective distance score between the target object image and the known object image based at least on the comparison of the digital pixel data of the target object image to the digital pixel data from the one or more known object images; (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph, Pg. 1539 § 2 - § 2.2, Pg. 1539 Figs. 3 & 4, Pg. 1540 § 3.1 ¶ 1, Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3 ¶ 1, Pg. 1541 Fig. 7, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) wherein, for a given known object image, a distance score is generated that represents a statistical similarity between the digital pixel data of the target object image and the digital pixel data of the given known object image; (Craparotta et al., Pg. 1539 § 2 - § 2.2, Pg. 1539 Fig. 4, Pg. 1540 § 3.1 ¶ 1, Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3 ¶ 1, Pg. 1541 Fig. 7, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1) wherein the known object images that are more similar to the target object image receive a lower distance score than a less similar known object image; (Craparotta et al., Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3 ¶ 1, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1) identifying, by the machine learning model, a set of similar object images that include the one or more known object images having a distance score below a threshold distance score; (Craparotta et al., Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) for each similar object image of the set of similar object images, retrieving object attributes including historical event data associated with the respective similar object image; (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph, Pg. 1541 § 3.1.3, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5) weighting, for each similar object image in the set of similar object images, the historical event data associated with the respective similar object image; (Craparotta et al., Pg. 1538 Left-Hand Colum Second-Full Paragraph, Pg. 1541 § 3.1.3, Pg. 1543 § 4.2, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) generating, by the machine learning model, the predicted characteristic model including a predicted characteristic for the target object represented in the target object image based at least on the weighted historical event data associated with the set of similar object images, (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph, Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5) wherein the predicted characteristic includes a predicted initial price for the target object; (Craparotta et al., Pg. 1540 § 3.1.1 - Pg. 1541 § 3.1.3 ¶ 1, Pg. 1540 Fig. 5, Pg. 1542 § 4 - § 4.1) and generating an electronic message with the predicted characteristic for the target object. (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph - Second-Full Paragraph, Pg. 1538 Fig. 2, Pg. 1540 § 3 - § 3.1, Pg. 1540 Fig. 5, Pg. 1541 § 3.1.3, Pg. 1542 § 4.1 - Pg. 1543 § 4.2, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) Craparotta et al. fail to disclose explicitly a computing system comprising at least one processor, a similarity score; wherein images that are more similar to the target image receive a greater similarity score than a less similar image; identifying a similarity score above a threshold similarity score; weighting, for each similar object image in the set of similar object images, the historical event data associated with the respective similar object image with the similarity score associated with the respective similar object image; wherein a greater similarity score causes the historical event data associated with the respective similar object image to have a greater influence on a predicted characteristic model than historical event data associated with a different similar object image having a lower similarity score; and transmitting the electronic message to a remote computer. Pertaining to analogous art, Tibau-Puig et al. disclose a method performed by a computing system comprising at least one processor, (Tibau-Puig et al., Abstract, Figs. 1A - 2, 3B, 4 & 7, Pg. 2 ¶ 0021 - 0024, Pg. 4 ¶ 0032 - 0033 and 0038, Pg. 6 ¶ 0049 - 0052, Pg. 7 ¶ 0056 - 0057) the method comprising: inputting, to a machine learning model, a target object image in digital form that represents a target object; (Tibau-Puig et al., Abstract, Figs. 1A - 4, Pg. 1 ¶ 0016 - Pg. 2 ¶ 0017, Pg. 2 ¶ 0019 - 0024, Pg. 3 ¶ 0030 - 0031) extracting, by the machine learning model executed by the one or more processors, digital pixel data of the target object image; (Tibau-Puig et al., Figs. 3B, 4 & 7, Pg. 2 ¶ 0021 - 0024, Pg. 4 ¶ 0033 - 0034, Pg. 5 ¶ 0039 and 0043, Pg. 6 ¶ 0049 - 0053) processing, by the machine learning model, the digital pixel data of the target object image by pixel analysis and comparing to pixel arrays of digital pixel data of a group of known object images; (Tibau-Puig et al., Figs. 3A - 4, Pg. 1 ¶ 0016, Pg. 2 ¶ 0020 - 0024, Pg. 3 ¶ 0030 - 0031, Pg. 5 ¶ 0039 - 0040 and 0042 - 0043) performing a comparison, by the machine learning model, of the digital pixel data of the target object image to digital pixel data associated with a group of known object images; (Tibau-Puig et al., Figs. 3A - 4, Pg. 1 ¶ 0016, Pg. 2 ¶ 0020 - 0024, Pg. 3 ¶ 0030 - 0031, Pg. 5 ¶ 0039 - 0040 and 0042 - 0043) identifying, by the machine learning model, a set of similar object images; (Tibau-Puig et al., Figs. 2 - 4, Pg. 1 ¶ 0016, Pg. 2 ¶ 0020 - 0024, Pg. 3 ¶ 0030 - 0031, Pg. 5 ¶ 0039 - 0040 and 0042 - 0043) for each similar object image of the set of similar object images, retrieving object attributes including historical event data associated with the respective similar object image; (Tibau-Puig et al., Abstract, Figs. 1A - 6, Pg. 2 ¶ 0019 and 0021, Pg. 2 ¶ 0023 - Pg. 3 ¶ 0026, Pg. 3 ¶ 0030 - 0031, Pg. 4 ¶ 0035 - 0036, Pg. 5 ¶ 0040 and 0044, Pg. 6 ¶ 0052 - 0053) generating, by the machine learning model, the predicted characteristic model including a predicted characteristic for the target object represented in the target object image based at least on the historical event data associated with the set of similar object images, (Tibau-Puig et al., Abstract, Figs. 1A - 6, Pg. 2 ¶ 0019 and 0021, Pg. 2 ¶ 0023 - Pg. 3 ¶ 0026, Pg. 3 ¶ 0030 - 0031, Pg. 4 ¶ 0035 - 0036, Pg. 5 ¶ 0040 and 0044, Pg. 6 ¶ 0052 - 0053) wherein the predicted characteristic includes a predicted initial price for the target object; (Tibau-Puig et al., Figs. 3B - 6, Pg. 2 ¶ 0019 - 0020 and 0023 - 0024, Pg. 3 ¶ 0026 and 0030 - 0031, Pg. 4 ¶ 0034, Pg. 5 ¶ 0040 and 0044) generating an electronic message with the predicted characteristic for the target object; (Tibau-Puig et al., Abstract, Figs. 1A & 3B - 5, Pg. 2 ¶ 0019, Pg. 2 ¶ 0023 - Pg. 3 ¶ 0027, Pg. 4 ¶ 0032 - 0034 and 0038, Pg. 5 ¶ 0040 - 0042, Pg. 6 ¶ 0054) and transmitting the electronic message to a remote computer. (Tibau-Puig et al., Figs. 1A, 1B, 3B & 4, Pg. 2 ¶ 0024 - Pg. 3 ¶ 0027, Pg. 4 ¶ 0032 - 0034 and 0038, Pg. 6 ¶ 0050 - 0051 and 0053 - 0054) Tibau-Puig et al. fail to disclose explicitly a similarity score; wherein images that are more similar to the target image receive a greater similarity score than a less similar image; identifying a similarity score above a threshold similarity score; weighting, for each similar object image in the set of similar object images, the historical event data associated with the respective similar object image with the similarity score associated with the respective similar object image; and wherein a greater similarity score causes the historical event data associated with the respective similar object image to have a greater influence on a predicted characteristic model than historical event data associated with a different similar object image having a lower similarity score. Pertaining to analogous art, Ekambaram et al. disclose weighting, for each similar object image in the set of similar object images, the historical event data associated with the respective similar object image with the distance score associated with the respective similar object image; (Ekambaram et al., Pg. 3112 § 3.1 - § 3.1.2 [“The estimator
ℇ
a produces a sales time-series estimate by performing weighted aggregation on time-series of k nearest neighbors”]) wherein a lower distance score (greater similarity score) causes the historical event data associated with the respective similar object image to have a greater influence on a predicted characteristic model than historical event data associated with a different similar object image having a greater distance score (lower similarity score); (Ekambaram et al., Pg. 3112 § 3.1 - § 3.1.2 [“The estimator
ℇ
a produces a sales time-series estimate by performing weighted aggregation on time-series of k nearest neighbors”, “represent the product features as a vector embedding. The vector embedding defines a proper distance metric. It is expected that similar products are close to each other compared to dissimilar products in this embedding space”, “X𝑝, and
Y
𝑝 represents the image/unstructured data and sales time-series attributed to product p. 𝑑 (Φ(X𝑝𝑖), Φ(X𝑝𝑗)) is the distance metric between product p𝑖, and p𝑗 in the embedding space” and “similar to attribute-based KNN, a KNN estimator ƐI is defined as
PNG
media_image1.png
60
420
media_image1.png
Greyscale
”. The Examiner asserts that one of ordinary skill in the art would understand that lower distances correspond to greater degrees of similarity, i.e., a lower distance score corresponds to a greater similarity score and vice-a-versa.]) and generating, by the machine learning model, the predicted characteristic model including a predicted characteristic for the target object represented in the target object image based at least on the weighted historical event data associated with the set of similar object images, (Ekambaram et al., Pg. 3112 § 3.1 - § 3.1.2) wherein the predicted characteristic includes a predicted initial price for the target object. (Ekambaram et al., Pg. 3111 § 1.1 ¶ 1, Pg. 3112 § 3.1, Pg. 3113 § 3.2.1 ¶ 1 - 2, Pg. 3115 § 4 ¶ 1) Ekambaram et al. fail to disclose expressly a similarity score; wherein images that are more similar to the target image receive a greater similarity score than a less similar image; and identifying a similarity score above a threshold similarity score. Pertaining to analogous art, Ikeda et al. disclose generating a similarity score between the target object image and the known object image based at least on the comparison of the digital pixel data of the target object image to the digital pixel data from the one or more known object images; (Ikeda et al., Figs. 2 & 8, Pg. 3 ¶ 0040 - 0043, Pg. 4 ¶ 0046 - 0048) wherein, for a given known object image, a similarity score is generated that represents a statistical similarity between the digital pixel data of the target object image and the digital pixel data of the given known object image; (Ikeda et al., Figs. 2 & 8, Pg. 3 ¶ 0040 - 0043, Pg. 4 ¶ 0046 - 0048) wherein the known object images that are more similar to the target object image receive a greater similarity score than a less similar known object image; (Ikeda et al., Pg. 4 ¶ 0046 - 0048, Pg. 5 ¶ 0058 - 0059 [“assuming that the reciprocal of the distance is the similarity degree, the smaller the distance, the larger the similarity degree”]) and identifying a similarity score above a threshold similarity score. (Ikeda et al., Pg. 4 ¶ 0046 - 0048, Pg. 5 ¶ 0058 - 0059 [“determination section 54 compares the similarity degree calculated by the similarity degree calculation section 52 with a predetermined threshold value. If the similarity degree is larger than the threshold value, the determination section 54 determines that the object being recognized is identical with the reference object.”]) Craparotta et al. and Tibau-Puig et al. are combinable because they are both directed towards image processing systems that analyze images of products to estimate prices and/or demand for the products. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Craparotta et al. with the teachings of Tibau-Puig et al. These modifications would have been prompted in order to enhance the base device of Craparotta et al. with the well-known and applicable techniques Tibau-Puig et al. applied to a comparable device. Utilizing a computing system comprising at least one processor to perform a method, as taught by Tibau-Puig et al., would enhance the base device of Craparotta et al. by ensuring that its operations are carried out accurately and efficiently at high-computational speed on computer architecture and by facilitating widespread distribution of the base device of Craparotta et al. to millions of potential end-users with access to a computer system. Additionally, transmitting the electronic message to a remote computer, as taught by Tibau-Puig et al., would enhance the base device of Craparotta et al. by enabling results obtained by the base device of Craparotta et al., the electronic message, to be easily shared between different users and/or locations and/or by allowing for end-users with devices having low computational processing power to offload most or all of the computationally intensive processing tasks to a remotely located and computationally powerful computing system so as to increase the overall operational speed of the base device of Craparotta et al. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that a computing system comprising at least one processor would be utilized to implement the base device of Craparotta et al. so as to ensure that its operations are carried out accurately and efficiently at high-computational speed on computer architecture and in that the electronic message would be transmitted to a remote computer so as to enable the electronic message to be easily shared between different users and/or locations and/or to allow for end-users with devices having low computational processing power to offload most or all of the computationally intensive processing tasks to a remotely located and computationally powerful computing system so as to increase the overall operational speed of the base device of Craparotta et al. In addition, Craparotta et al. in view of Tibau-Puig et al. and Ekambaram et al. are combinable because they are all directed towards image processing systems that analyze images of products to estimate prices and/or demand for the products. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. with the teachings of Ekambaram et al. This modification would have been prompted in order to enhance the combined base device of Craparotta et al. in view of Tibau-Puig et al. with the well-known and applicable technique Ekambaram et al. applied to a comparable device. Weighting the historical event data for each similar object image with the distance score associated with each of the similar object images when generating the predicted characteristic model for the target object, as taught by Ekambaram et al., would enhance the combined base device by helping improve its ability to accurately and reliably predict the sales profile of an imaged product since sales profiles for products used in predicting the sales profile of the imaged product would be weighted so that sales profiles for products that are the most similar to the imaged product have more of an influence on the predicted sales profile than sales profiles of less similar products thereby allowing for products which are more comparable to the imaged product to have more of an impact on the predicted sales profile of the imaged product. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that historical event data for each similar object image would be weighted with the distance score associated with each of the similar object images when generating the predicted characteristic model for the target object so as to allow for objects which are more comparable to the target object than others to have more of an influence on the predicted characteristic model for the target object than less comparable objects in order to help improve the ability of the combined base device to accurately and reliably generate the predicted characteristic model for the target object. Additionally, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. and Ikeda et al. are combinable because they are all directed towards image processing systems that classify images of objects and, similar to Craparotta et al. and Ekambaram et al., Ikeda et al. also identify images that are similar to an input image. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. with the teachings of Ikeda et al. This modification would have been prompted in order to substitute the distance score of the combined base device for the similarity degree of Ikeda et al. The similarity degree of Ikeda et al. could be substituted in place of the distance score of the combined base device utilizing well-known techniques in the art and would likely yield predictable results, in that in the combination the similarity degree of Ikeda et al. would be utilized to define the similarity between images/products, identify similar images and weight the historical event data. Furthermore, modification would have been prompted by the teachings and suggestions of Ekambaram et al. that their vector embedding defines a proper distance metric, that it is expected that similar products are close to each other compared to dissimilar products in their embedding space and that the similarity between products can be defined based on the cosine distance between their image embeddings, see at least pages 3112 - 3113 section 3.1 and page 3116 section 4.2 paragraph 1 of Ekambaram et al. Moreover, this modification would have been prompted by the teachings and suggestions of Ikeda et al. that the reciprocal of a distance between images can be treated as a similarity degree between images, see at least page 4 paragraphs 0046 - 0048 of Ikeda et al. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that the similarity degree of Ikeda et al., the reciprocal of distance, would be used in place of the distance score of the combined base device and utilized by the combined base device to define the similarity between images/products, identify similar images and weight the historical event data. Therefore, it would have been obvious to combine Craparotta et al. with Tibau-Puig et al., Ekambaram et al. and Ikeda et al. to obtain the invention as specified in claim 1.
- With regards to claim 2, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. disclose the method of claim 1, wherein the machine learning model is configured with a neural network model that is trained to classify images of objects. (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph - Pg. 1539 § 2.2, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) Craparotta et al. fail to disclose explicitly adjusting, by the machine learning model, the predicted characteristic for the target object with additional object attributes from the set of similar object images prior to generating the electronic message. Pertaining to analogous art, Tibau-Puig et al. disclose wherein the machine learning model is configured with a neural network model that is trained to classify images of objects; (Tibau-Puig et al., Pg. 1 ¶ 0016 - Pg. 2 ¶ 0017, Pg. 2 ¶ 0021 - 0024, Pg. 3 ¶ 0030 - 0031, Pg. 5 ¶ 0042 - 0043) and adjusting, by the machine learning model, the predicted characteristic for the target object with additional object attributes from the set of similar object images prior to generating the electronic message. (Tibau-Puig et al., Pg. 2 ¶ 0019 - 0021 and 0023 - 0024, Pg. 3 ¶ 0026 - 0027 and 0030 - 0031, Pg. 4 ¶ 0034 - 0037) It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with additional teachings of Tibau-Puig et al. This modification would have been prompted in order to enhance the combined base device of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the well-known and applicable technique Tibau-Puig et al. applied to a comparable device. Adjusting, by the machine learning model, the predicted characteristic for the target object with additional object attributes from the set of similar object images, as taught by Tibau-Puig et al., would enhance the combined base device by helping ensure that it is able to predict the sales profile of an imaged product as accurately and reliably as possible since additional attributes of products identified as being similar to the image product would be utilized to further adjust and refine the sales profile predicted for the imaged product thereby allowing for the combined base device to take into account similarities and/or differences between the imaged product and similar products identified when generating the sales profile for the imaged product. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that the predicted characteristic for the target object would be adjusted, by the machine learning model, with additional object attributes from the set of similar object images so as to allow for the combined base device to take into account and adjust for similarities and/or differences between attributes of an imaged product and attributes of identified similar products when generating the predicted characteristic, sales profile, for the imaged product. Therefore, it would have been obvious to combine Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with additional teachings of Tibau-Puig et al. to obtain the invention as specified in claim 2.
- With regards to claim 3, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. disclose the method of claim 1, further comprising: analyzing, by the machine learning model, the digital pixel data of the target object image including object-based image analysis to group pixels to identify the target object in the target object image.(Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph - Pg. 1539 § 2.2, Pg. 1539 Fig. 3, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1541 Fig. 7, Pg. 1542 § 4 - § 4.1 ¶ 2, Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) In addition, analogous art Tibau-Puig et al. disclose analyzing, by the machine learning model, the digital pixel data of the target object image including object-based image analysis to group pixels to identify the target object in the target object image. (Tibau-Puig et al., Pg. 1 ¶ 0016 - Pg. 2 ¶ 0017, Pg. 2 ¶ 0021 - 0024, Pg. 5 ¶ 0039 and 0042 - 0043)
- With regards to claim 5, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. disclose the method of claim 1, wherein the target object is a target product, (Craparotta et al., Pg. 1537 Abstract, Pg. 1538 Right-Hand Column First-Full Paragraph, Pg. 1538 Fig. 2, Pg. 1540 § 3 - § 3.1, Pg. 1540 Fig. 5, Pg. 1542 § 4 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) and wherein generating the predicted characteristic further comprises: configuring the predicted characteristic model to generate the predicted characteristic as a predicted demand for the target product (i), at a first store (S) which is based on a function, (Craparotta et al., Pg. 1537 Abstract, Pg. 1538 Left-Hand Column Second-Full Paragraph - Right-Hand Column First-Full Paragraph, Pg. 1540 § 3 - § 3.1.1, Pg. 1540 Figs. 5 & 6, Pg. 1541 § 3.1.3, Pg. 1542 § 4 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5) where the target product (i) is associated with the target object image which is determined to be similar to the known products labeled as (j1, j2, ... jk) that are represented in the one or more known object images that are identified as the set of similar object images, (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph - Second-Full Paragraph, Pg. 1538 Fig. 2, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1542 § 4 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) and sim_i_j1, sim_i_j2,... and sim_i_jk represent distance scores generated for each of the one or more known products (j1, j2, ... jk) from the set of similar object images as compared to the target product (i). (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph, Pg. 1539 § 2 - § 2.2, Pg. 1539 Figs. 3 & 4, Pg. 1540 § 3.1 ¶ 1, Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3 ¶ 1, Pg. 1541 Fig. 7, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16 [The Examiner asserts that, in the proposed combination of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al., the similarity scores of Ikeda et al. would be utilized in place of distance scores.]) Craparotta et al. fail to disclose explicitly a function (sim_i_j1 * demand for a known product j1), (sim_i_j2 * demand for a known product j2), ... (sim_i_jk * demand for a known product jk) and similarity scores. Pertaining to analogous art, Ekambaram et al. disclose wherein generating the predicted characteristic further comprises: configuring the predicted characteristic model to generate the predicted characteristic as a predicted demand for the target product (i), at a first store (S) which is based on a function (sim_i_j1 * demand for a known product j1), (sim_i_j2 * demand for a known product j2), ... (sim_i_jk * demand for a known product jk), (Ekambaram et al., Pg. 3112 § 3 - § 3.1.2 [“The estimator
ℇ
a produces a sales time-series estimate by performing weighted aggregation on time-series of k nearest neighbors”, “
X
𝑝, and
Y
𝑝 represents the image/unstructured data and sales time-series attributed to product p” and “𝑑 (Φ(
X
𝑝𝑖), Φ(
Y
𝑝𝑗)) is the distance metric between product p𝑖, and p𝑗 in the embedding space”]) where the target product (i) is associated with the target object image which is determined to be similar to the known products labeled as (j1, j2, ... jk) that are represented in the one or more known object images that are identified as the set of similar object images, (Ekambaram et al., Pg. 3111 § 1.1, Pg. 3112 § 3 - § 3.1.2) and sim_i_j1, sim_i_j2,... and sim_i_jk represent distance scores generated for each of the one or more known products (j1, j2, ... jk) from the set of similar object images as compared to the target product (i). (Ekambaram et al., Pg. 3111 § 1.1, Pg. 3112 § 3 - § 3.1.2 [“𝑑 (Φ(
X
𝑝𝑖), Φ(
X
𝑝𝑗)) is the distance metric between product p𝑖, and p𝑗 in the embedding space.” The Examiner asserts that, in the proposed combination of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al., the similarity scores of Ikeda et al. would be utilized in place of the distance scores.]) Ekambaram et al. fail to disclose expressly similarity scores. Pertaining to analogous art, Ikeda et al. disclose where sim_i_j1, sim_i_j2,... and sim_i_jk represent the similarity scores generated for each of the one or more known products (j1, j2, ... jk) from the set of similar object images as compared to the target product (i). (Ikeda et al., Figs. 2 & 8, Pg. 3 ¶ 0040 - 0043, Pg. 4 ¶ 0046 - 0048 [The Examiner asserts that, in the proposed combination of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al., Ikeda et al. would generate similarity scores for each of the one or more known products (j1, j2, ... jk) from the set of similar object images as compared to the target product (i).])
- With regards to claim 8, Craparotta et al. disclose instructions, (Craparotta et al., Pg. 1537 Abstract, Pg. 1540 § 3 ¶ 1, Pg. 1540 Fig. 5, Pgs. 1544 - 1545 § 5) wherein the instructions comprise: input, to a machine learning model, a target product image in digital form that represents a target product; (Craparotta et al., Pg. 1538 Fig. 2, Pg. 1539 § 2.1 - § 2.2, Pg. 1539 Figs. 3 & 4, Pg. 1540 Fig. 5, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1541 Fig. 7, Pg. 1542 § 4.1 ¶ 1 - 2, Pg. 1543 Fig. 11, Pg. 1544 § 5, Pg. 1545 Fig. 16) identify, by the machine learning model, a set of similar product images (Craparotta et al., Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) by comparing digital pixel data of the target product image to digital pixel data of known product images (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph - Pg. 1539 § 2.2, Pg. 1538 Fig. 2, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1541 Fig. 7, Pg. 1542 § 4 - § 4.1 ¶ 2, Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pg. 1544 § 5 ¶ 1, Pg. 1545 Fig. 16) and generating a distance score between the target product image and each similar product image of the set of similar product images; (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph, Pg. 1539 § 2 - § 2.2, Pg. 1539 Figs. 3 & 4, Pg. 1540 § 3.1 ¶ 1, Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3 ¶ 1, Pg. 1541 Fig. 7, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) wherein the known product images that are more similar to the target product image receive a lower distance score than a less similar known product image; (Craparotta et al., Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3 ¶ 1, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1) for each similar product image of the set of similar product images, retrieve product attributes including historical event data associated with the respective similar product image; (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph, Pg. 1541 § 3.1.3, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5) weight, for each similar product image in the set of similar product images, the historical event data associated with the respective similar product image; (Craparotta et al., Pg. 1538 Left-Hand Colum Second-Full Paragraph, Pg. 1541 § 3.1.3, Pg. 1543 § 4.2, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) and generate, by the machine learning model, the predicted characteristic for the target product represented in the target product image based at least on the weighted historical event data associated with the set of similar product images, (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph, Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5) the predicted characteristic including a predicted initial price for the target product. (Craparotta et al., Pg. 1540 § 3.1.1 - Pg. 1541 § 3.1.3 ¶ 1, Pg. 1540 Fig. 5, Pg. 1542 § 4 - § 4.1) Craparotta et al. fail to disclose explicitly a non-transitory computer-readable medium storing computer-executable instructions that, when executed by a computer including a processor, cause the computer to perform functions configured by the computer-executable instructions; a similarity score; wherein images that are more similar to the target image receive a greater similarity score than a less similar known image; weighting, for each similar product image in the set of similar product images, the historical event data associated with the respective similar product image with the similarity score associated with the respective similar product image; wherein a greater similarity score causes the historical event data associated with the respective similar product image to have a greater influence on a predicted characteristic model than historical event data associated with a different similar product image having a lower similarity score; and generating instructions to cause a database to assign the predicted characteristic including the predicted initial price to one or more data records associated with the target product. Pertaining to analogous art, Tibau-Puig et al. disclose a non-transitory computer-readable medium storing computer-executable instructions that, when executed by a computer including a processor, cause the computer to perform functions configured by the computer-executable instructions, (Tibau-Puig et al., Fig. 7, Pg. 2 ¶ 0021 and 0024, Pg. 4 ¶ 0032 - 0034 and 0038, Pg. 6 ¶ 0049 - 0052, Pg. 7 ¶ 0056 - 0057) wherein the instructions cause the computer to: input, to a machine learning model, a target product image in digital form that represents a target product; (Tibau-Puig et al., Abstract, Figs. 1A - 4, Pg. 1 ¶ 0016 - Pg. 2 ¶ 0017, Pg. 2 ¶ 0019 - 0024, Pg. 3 ¶ 0030 - 0031) identify, by the machine learning model, a set of similar product images (Tibau-Puig et al., Figs. 2 - 4, Pg. 1 ¶ 0016, Pg. 2 ¶ 0020 - 0024, Pg. 3 ¶ 0030 - 0031, Pg. 5 ¶ 0039 - 0040 and 0042 - 0043) by comparing digital pixel data of the target product image to digital pixel data of known product images; (Tibau-Puig et al., Figs. 3A - 4, Pg. 1 ¶ 0016, Pg. 2 ¶ 0020 - 0024, Pg. 3 ¶ 0030 - 0031, Pg. 5 ¶ 0039 - 0040 and 0042 - 0043) for each similar product image of the set of similar product images, retrieve product attributes including historical event data associated with the respective similar product image; (Tibau-Puig et al., Abstract, Figs. 1A - 6, Pg. 2 ¶ 0019 and 0021, Pg. 2 ¶ 0023 - Pg. 3 ¶ 0026, Pg. 3 ¶ 0030 - 0031, Pg. 4 ¶ 0035 - 0036, Pg. 5 ¶ 0040 and 0044, Pg. 6 ¶ 0052 - 0053) generate, by the machine learning model, the predicted characteristic for the target product represented in the target product image based at least on the historical event data associated with the set of similar product images, (Tibau-Puig et al., Abstract, Figs. 1A - 6, Pg. 2 ¶ 0019 and 0021, Pg. 2 ¶ 0023 - Pg. 3 ¶ 0026, Pg. 3 ¶ 0030 - 0031, Pg. 4 ¶ 0035 - 0036, Pg. 5 ¶ 0040 and 0044, Pg. 6 ¶ 0052 - 0053) the predicted characteristic including a predicted initial price for the target product; (Tibau-Puig et al., Figs. 3B - 6, Pg. 2 ¶ 0019 - 0020 and 0023 - 0024, Pg. 3 ¶ 0026 and 0030 - 0031, Pg. 4 ¶ 0034, Pg. 5 ¶ 0040 and 0044) and generate instructions to cause a database to assign the predicted characteristic including the predicted initial price to one or more data records associated with the target product. (Tibau-Puig et al., Figs. 1B & 3B - 7, Pg. 2 ¶ 0019 - 0021 and 0023 - 0024, Pg. 3 ¶ 0026, 0028 and 0030 - 0031, Pg. 4 ¶ 0034 and 0038, Pg. 5 ¶ 0040, Pg. 5 ¶ 0043 - Pg. 6 ¶ 0048, Pg. 6 ¶ 0052, Pg. 7 ¶ 0055 - 0057) Tibau-Puig et al. fail to disclose explicitly a similarity score; wherein images that are more similar to the target image receive a greater similarity score than a less similar known image; weighting, for each similar product image in the set of similar product images, the historical event data associated with the respective similar product image with the similarity score associated with the respective similar product image; and wherein a greater similarity score causes the historical event data associated with the respective similar product image to have a greater influence on a predicted characteristic model than historical event data associated with a different similar product image having a lower similarity score. Pertaining to analogous art, Ekambaram et al. disclose weighting, for each similar product image in the set of similar product images, the historical event data associated with the respective similar product image with the distance score associated with the respective similar product image; (Ekambaram et al., Pg. 3112 § 3.1 - § 3.1.2 [“The estimator
ℇ
a produces a sales time-series estimate by performing weighted aggregation on time-series of k nearest neighbors”]) wherein a lower distance score (greater similarity score) causes the historical event data associated with the respective similar product image to have a greater influence on a predicted characteristic model than historical event data associated with a different similar product image having a greater distance score (lower similarity score); (Ekambaram et al., Pg. 3112 § 3.1 - § 3.1.2 [“The estimator
ℇ
a produces a sales time-series estimate by performing weighted aggregation on time-series of k nearest neighbors”, “represent the product features as a vector embedding. The vector embedding defines a proper distance metric. It is expected that similar products are close to each other compared to dissimilar products in this embedding space”, “X𝑝, and
Y
𝑝 represents the image/unstructured data and sales time-series attributed to product p. 𝑑 (Φ(X𝑝𝑖), Φ(X𝑝𝑗)) is the distance metric between product p𝑖, and p𝑗 in the embedding space” and “similar to attribute-based KNN, a KNN estimator ƐI is defined as
PNG
media_image1.png
60
420
media_image1.png
Greyscale
”. The Examiner asserts that one of ordinary skill in the art would understand that lower distances correspond to greater degrees of similarity, i.e., a lower distance score corresponds to a greater similarity score and vice-a-versa.]) and generating, by the machine learning model, the predicted characteristic for the target product represented in the target product image based at least on the weighted historical event data associated with the set of similar product images, (Ekambaram et al., Pg. 3112 § 3.1 - § 3.1.2) the predicted characteristic including a predicted initial price for the target product. (Ekambaram et al., Pg. 3111 § 1.1 ¶ 1, Pg. 3112 § 3.1, Pg. 3113 § 3.2.1 ¶ 1 - 2, Pg. 3115 § 4 ¶ 1) Ekambaram et al. fail to disclose expressly a similarity score; wherein images that are more similar to the target image receive a greater similarity score than a less similar known image. Pertaining to analogous art, Ikeda et al. disclose generating a similarity score between the target product image and each similar product image of the set of similar product images; (Ikeda et al., Figs. 2 & 8, Pg. 3 ¶ 0040 - 0043, Pg. 4 ¶ 0046 - 0048) wherein the known product images that are more similar to the target product image receive a greater similarity score than a less similar known product image. (Ikeda et al., Pg. 4 ¶ 0046 - 0048, Pg. 5 ¶ 0058 - 0059 [“assuming that the reciprocal of the distance is the similarity degree, the smaller the distance, the larger the similarity degree”]) Craparotta et al. and Tibau-Puig et al. are combinable because they are both directed towards image processing systems that analyze images of products to estimate prices and/or demand for the products. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Craparotta et al. with the teachings of Tibau-Puig et al. These modifications would have been prompted in order to enhance the base device of Craparotta et al. with the well-known and applicable techniques Tibau-Puig et al. applied to a comparable device. Utilizing a non-transitory computer-readable medium storing computer-executable instructions that, when executed by a computer including a processor, cause the computer to perform functions, as taught by Tibau-Puig et al., to implement operations of the base device of Craparotta et al. would enhance the base device of Craparotta et al. by ensuring that its operations are carried out accurately and efficiently at high-computational speed on computer architecture, by facilitating widespread distribution of the base device of Craparotta et al. to millions of potential end-users with access to a computer and by simplifying the process of making revisions, modifications and/or updates to the operations of the base device of Craparotta et al. Additionally, causing a database to assign the predicted characteristic including the predicted initial price to one or more data records associated with the target product, as taught by Tibau-Puig et al., would enhance the base device of Craparotta et al. by allowing for predicted characteristics of target products to be preserved and readily available for future reference, analysis, and/or retrieval so as to provide end-users with the ability to subsequently utilize the predicted characteristics of target products for any of a variety reasons and/or applications. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that a non-transitory computer-readable medium storing computer-executable instructions that, when executed by a computer including a processor, cause the computer to perform functions would be utilized to implement the base device of Craparotta et al. so as to ensure that its operations are carried out accurately and efficiently at high-computational speed on computer architecture, to facilitate widespread distribution of the base device of Craparotta et al. to millions of potential end-users with access to a computer and to simplify the process of making revisions, modifications and/or updates to the operations of the base device of Craparotta et al. and in that a database would assign the predicted characteristic to data records associated with the target product in order to allow for the predicted characteristic of the target product to be preserved and readily available for future reference, analysis, and/or retrieval so as to provide end-users with the ability to subsequently utilize the predicted characteristic of the target product for any of a variety reasons. In addition, Craparotta et al. in view of Tibau-Puig et al. and Ekambaram et al. are combinable because they are all directed towards image processing systems that analyze images of products to estimate prices and/or demand for the products. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. with the teachings of Ekambaram et al. This modification would have been prompted in order to enhance the combined base device of Craparotta et al. in view of Tibau-Puig et al. with the well-known and applicable technique Ekambaram et al. applied to a comparable device. Weighting the historical event data for each similar object image with the distance score associated with each of the similar object images when generating the predicted characteristic model for the target object, as taught by Ekambaram et al., would enhance the combined base device by helping improve its ability to accurately and reliably predict the sales profile of an imaged product since sales profiles for products used in predicting the sales profile of the imaged product would be weighted so that sales profiles for products that are the most similar to the imaged product have more of an influence on the predicted sales profile than sales profiles of less similar products thereby allowing for products which are more comparable to the imaged product to have more of an impact on the predicted sales profile of the imaged product. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that historical event data for each similar object image would be weighted with the distance score associated with each of the similar object images when generating the predicted characteristic model for the target object so as to allow for objects which are more comparable to the target object than others to have more of an influence on the predicted characteristic model for the target object than less comparable objects in order to help improve the ability of the combined base device to accurately and reliably generate the predicted characteristic model for the target object. Additionally, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. and Ikeda et al. are combinable because they are all directed towards image processing systems that classify images of objects and, similar to Craparotta et al. and Ekambaram et al., Ikeda et al. also identify images that are similar to an input image. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. with the teachings of Ikeda et al. This modification would have been prompted in order to substitute the distance score of the combined base device for the similarity degree of Ikeda et al. The similarity degree of Ikeda et al. could be substituted in place of the distance score of the combined base device utilizing well-known techniques in the art and would likely yield predictable results, in that in the combination the similarity degree of Ikeda et al. would be utilized to define the similarity between images/products, identify similar images and weight the historical event data. Furthermore, modification would have been prompted by the teachings and suggestions of Ekambaram et al. that their vector embedding defines a proper distance metric, that it is expected that similar products are close to each other compared to dissimilar products in their embedding space and that the similarity between products can be defined based on the cosine distance between their image embeddings, see at least pages 3112 - 3113 section 3.1 and page 3116 section 4.2 paragraph 1 of Ekambaram et al. Moreover, this modification would have been prompted by the teachings and suggestions of Ikeda et al. that the reciprocal of a distance between images can be treated as a similarity degree between images, see at least page 4 paragraphs 0046 - 0048 of Ikeda et al. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that the similarity degree of Ikeda et al., the reciprocal of distance, would be used in place of the distance score of the combined base device and utilized by the combined base device to define the similarity between images/products, identify similar images and weight the historical event data. Therefore, it would have been obvious to combine Craparotta et al. with Tibau-Puig et al., Ekambaram et al. and Ikeda et al. to obtain the invention as specified in claim 8.
- With regards to claim 9, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. disclose the non-transitory computer-readable medium of claim 8. Craparotta et al. fail to disclose explicitly instructions that, when executed by at least the processor, cause the processor to: adjust the predicted characteristic for the target product with additional product attributes from the set of similar product images. Pertaining to analogous art, Tibau-Puig et al. disclose instructions that, when executed by at least the processor, cause the processor to: adjust the predicted characteristic for the target product with additional product attributes from the set of similar product images. (Tibau-Puig et al., Pg. 2 ¶ 0019 - 0021 and 0023 - 0024, Pg. 3 ¶ 0026 - 0027 and 0030 - 0031, Pg. 4 ¶ 0034 - 0037) It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with additional teachings of Tibau-Puig et al. This modification would have been prompted in order to enhance the combined base device of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the well-known and applicable technique Tibau-Puig et al. applied to a comparable device. Adjusting the predicted characteristic for the target product with additional product attributes from the set of similar product images, as taught by Tibau-Puig et al., would enhance the combined base device by helping ensure that it is able to predict the sales profile of an imaged product as accurately and reliably as possible since additional attributes of products identified as being similar to the image product would be utilized to further adjust and refine the sales profile predicted for the imaged product thereby allowing for the combined base device to take into account similarities and/or differences between the imaged product and similar products identified when generating the sales profile for the imaged product. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that the predicted characteristic for the target product would be adjusted with additional product attributes from the set of similar product images so as to allow for the combined base device to take into account and adjust for similarities and/or differences between attributes of an imaged product and attributes of identified similar products when generating the predicted characteristic, sales profile, for the imaged product. Therefore, it would have been obvious to combine Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with additional teachings of Tibau-Puig et al. to obtain the invention as specified in claim 9.
- With regards to claim 10, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. disclose the non-transitory computer-readable medium of claim 8, further comprising instructions that, when executed by at least the processor, cause the processor to: analyze, by the machine learning model, the digital pixel data of the target product image including object-based image analysis to group pixels to identify the target product in the target product image. (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph - Pg. 1539 § 2.2, Pg. 1539 Fig. 3, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1541 Fig. 7, Pg. 1542 § 4 - § 4.1 ¶ 2, Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) In addition, analogous art Tibau-Puig et al. disclose instructions that, when executed by at least the processor, cause the processor to: analyze, by the machine learning model, the digital pixel data of the target product image including object-based image analysis to group pixels to identify the target product in the target product image. (Tibau-Puig et al., Pg. 1 ¶ 0016 - Pg. 2 ¶ 0017, Pg. 2 ¶ 0021 - 0024, Pg. 5 ¶ 0039 and 0042 - 0043)
- With regards to claim 12, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. disclose the non-transitory computer-readable medium of claim 8, further comprising instructions that, when executed by at least the processor, cause the processor to: generate the predicted characteristic as a predicted demand for the target product (i) at a first store (S) which is based on a function, (Craparotta et al., Pg. 1537 Abstract, Pg. 1538 Left-Hand Column Second-Full Paragraph - Right-Hand Column First-Full Paragraph, Pg. 1540 § 3 - § 3.1.1, Pg. 1540 Figs. 5 & 6, Pg. 1541 § 3.1.3, Pg. 1542 § 4 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5) where the target product (i) is associated with the target product image which is determined to be similar to the known products (j1, j2, ... jk) that are represented in the known known product images, (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph - Second-Full Paragraph, Pg. 1538 Fig. 2, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1542 § 4 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) and sim_i_j1, sim_i_j2,... and sim_i_Jk represent distance scores between the target product (i) and each of the known products (j1, j2, ... jk). (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph, Pg. 1539 § 2 - § 2.2, Pg. 1539 Figs. 3 & 4, Pg. 1540 § 3.1 ¶ 1, Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3 ¶ 1, Pg. 1541 Fig. 7, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16 [The Examiner asserts that, in the proposed combination of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al., the similarity scores of Ikeda et al. would be utilized in place of distance scores.]) Craparotta et al. fail to disclose explicitly a function (sim_i_j1 * demand for a known product j1), (sim_i_j2 * demand for a known product j2), ... (sim_i_jk * demand for a known product jk) and similarity scores. Pertaining to analogous art, Ekambaram et al. disclose generating the predicted characteristic as a predicted demand for the target product (i) at a first store (S) which is based on a function (sim_i_j1 * demand for a known product j1), (sim_i_j2 * demand for a known product j2), ... (sim_i_jk * demand for a known product jk), (Ekambaram et al., Pg. 3112 § 3 - § 3.1.2 [“The estimator
ℇ
a produces a sales time-series estimate by performing weighted aggregation on time-series of k nearest neighbors”, “
X
𝑝, and
Y
𝑝 represents the image/unstructured data and sales time-series attributed to product p” and “𝑑 (Φ(
X
𝑝𝑖), Φ(
Y
𝑝𝑗)) is the distance metric between product p𝑖, and p𝑗 in the embedding space”]) where the target product (i) is associated with the target product image which is determined to be similar to the known products (j1, j2, ... jk) that are represented in the known known product images, (Ekambaram et al., Pg. 3111 § 1.1, Pg. 3112 § 3 - § 3.1.2) and sim_i_j1, sim_i_j2,... and sim_i_Jk represent distance scores between the target product (i) and each of the known products (j1, j2, ... jk). (Ekambaram et al., Pg. 3111 § 1.1, Pg. 3112 § 3 - § 3.1.2 [“𝑑 (Φ(
X
𝑝𝑖), Φ(
X
𝑝𝑗)) is the distance metric between product p𝑖, and p𝑗 in the embedding space.” The Examiner asserts that, in the proposed combination of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al., the similarity scores of Ikeda et al. would be utilized in place of the distance scores.]) Ekambaram et al. fail to disclose expressly the similarity scores. Pertaining to analogous art, Ikeda et al. disclose where sim_i_j1, sim_i_j2,... and sim_i_Jk represent the similarity scores between the target product (i) and each of the known products (j1, j2, ... jk). (Ikeda et al., Figs. 2 & 8, Pg. 3 ¶ 0040 - 0043, Pg. 4 ¶ 0046 - 0048 [The Examiner asserts that, in the proposed combination of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al., Ikeda et al. would generate similarity scores for each of the one or more known products (j1, j2, ... jk) from the set of similar object images as compared to the target product (i).])
- With regards to claim 15, Craparotta et al. disclose a computing system, (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph - Pg. 1539 § 2.2, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1541 Fig. 7, Pg. 1542 § 4 - Pg. 1543 § 4.2 ¶ 1, Pgs. 1544 - 1545 § 5) to: input, to a machine learning model, a target product image in digital form that represents a target product; (Craparotta et al., Pg. 1538 Fig. 2, Pg. 1539 § 2.1 - § 2.2, Pg. 1539 Figs. 3 & 4, Pg. 1540 Fig. 5, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1541 Fig. 7, Pg. 1542 § 4.1 ¶ 1 - 2, Pg. 1543 Fig. 11, Pg. 1544 § 5, Pg. 1545 Fig. 16) perform a comparison, by the machine learning model, at least digital pixel data of the target product image to digital pixel data from a group of known product images; (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph - Pg. 1539 § 2.2, Pg. 1538 Fig. 2, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1541 Fig. 7, Pg. 1542 § 4 - § 4.1 ¶ 2, Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pg. 1544 § 5 ¶ 1, Pg. 1545 Fig. 16) generate, by the machine learning model, for each of one or more known product images in the group of known product images, a distance score between the target product image and one or more known product images from the group of known product images based at least on the comparison; (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph, Pg. 1539 § 2 - § 2.2, Pg. 1539 Figs. 3 & 4, Pg. 1540 § 3.1 ¶ 1, Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3 ¶ 1, Pg. 1541 Fig. 7, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) wherein the known product images that are more similar to the target product image receive a lower distance score than a less similar known product image; (Craparotta et al., Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3 ¶ 1, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1) identify, by the machine learning model, a set of similar product images based at least in part on the similarity score of the one or more known product images having a distance score below a threshold distance score; (Craparotta et al., Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) for each similar product image of the set of similar product images, retrieve product attributes including historical event data associated with the respective similar product image; (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph, Pg. 1541 § 3.1.3, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5) weight, for each similar product image in the set of similar product images, the historical event data associated with the respective similar product image; (Craparotta et al., Pg. 1538 Left-Hand Colum Second-Full Paragraph, Pg. 1541 § 3.1.3, Pg. 1543 § 4.2, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) and generate, by the machine learning model, the predicted characteristic model for the target product represented in the target product image based at least on the weighted historical event data associated with the set of similar product images, (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph, Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5) the predicted characteristic model including a predicted initial price for the target product. (Craparotta et al., Pg. 1540 § 3.1.1 - Pg. 1541 § 3.1.3 ¶ 1, Pg. 1540 Fig. 5, Pg. 1542 § 4 - § 4.1) Craparotta et al. fail to disclose explicitly a computing system, comprising: at least one processor connected to at least one memory; and a non-transitory computer readable medium including instructions stored thereon that, when executed by the at least one processor, cause the at least one processor to perform functions; a similarity score; wherein images that are more similar to the target image receive a greater similarity score than a less similar image; identifying a similarity score above a threshold similarity score; weighting, for each similar product image in the set of similar product images, the historical event data associated with the respective similar product image based on the similarity score associated with the respective similar product image; wherein a greater similarity score causes the historical event data associated with the respective similar product image to have a greater influence on a predicted characteristic model than historical event data associated with a different similar product image having a lower similarity score; and transmit instructions to cause a database to assign the predicted characteristic model including the predicted initial price to one or more data records associated with the target product. Pertaining to analogous art, Tibau-Puig et al. disclose a computing system, (Tibau-Puig et al., Abstract, Figs. 1A - 2, 3B, 4 & 7, Pg. 2 ¶ 0021 - 0024, Pg. 4 ¶ 0032 - 0033 and 0038, Pg. 6 ¶ 0049 - 0052, Pg. 7 ¶ 0056 - 0057) comprising: at least one processor connected to at least one memory; (Tibau-Puig et al., Fig. 7, Pg. 2 ¶ 0021 and 0024, Pg. 4 ¶ 0032 - 0033 and 0038, Pg. 6 ¶ 0049 - 0052, Pg. 7 ¶ 0056 - 0057) and a non-transitory computer readable medium including instructions stored thereon that, when executed by the at least one processor, cause the at least one processor (Tibau-Puig et al., Fig. 7, Pg. 2 ¶ 0021 and 0024, Pg. 4 ¶ 0032 - 0034 and 0038, Pg. 6 ¶ 0049 - 0052, Pg. 7 ¶ 0056 - 0057) to: input, to a machine learning model, a target product image in digital form that represents a target product; (Tibau-Puig et al., Abstract, Figs. 1A - 4, Pg. 1 ¶ 0016 - Pg. 2 ¶ 0017, Pg. 2 ¶ 0019 - 0024, Pg. 3 ¶ 0030 - 0031) perform a comparison, by the machine learning model, at least digital pixel data of the target product image to digital pixel data from a group of known product images; (Tibau-Puig et al., Figs. 3A - 4, Pg. 1 ¶ 0016, Pg. 2 ¶ 0020 - 0024, Pg. 3 ¶ 0030 - 0031, Pg. 5 ¶ 0039 - 0040 and 0042 - 0043) identify, by the machine learning model, a set of similar product images; (Tibau-Puig et al., Figs. 2 - 4, Pg. 1 ¶ 0016, Pg. 2 ¶ 0020 - 0024, Pg. 3 ¶ 0030 - 0031, Pg. 5 ¶ 0039 - 0040 and 0042 - 0043) for each similar product image of the set of similar product images, retrieve product attributes including historical event data associated with the respective similar product image; (Tibau-Puig et al., Abstract, Figs. 1A - 6, Pg. 2 ¶ 0019 and 0021, Pg. 2 ¶ 0023 - Pg. 3 ¶ 0026, Pg. 3 ¶ 0030 - 0031, Pg. 4 ¶ 0035 - 0036, Pg. 5 ¶ 0040 and 0044, Pg. 6 ¶ 0052 - 0053) generate, by the machine learning model, the predicted characteristic model for the target product represented in the target product image based at least on the historical event data associated with the set of similar product images, (Tibau-Puig et al., Abstract, Figs. 1A - 6, Pg. 2 ¶ 0019 and 0021, Pg. 2 ¶ 0023 - Pg. 3 ¶ 0026, Pg. 3 ¶ 0030 - 0031, Pg. 4 ¶ 0035 - 0036, Pg. 5 ¶ 0040 and 0044, Pg. 6 ¶ 0052 - 0053) the predicted characteristic model including a predicted initial price for the target product; (Tibau-Puig et al., Figs. 3B - 6, Pg. 2 ¶ 0019 - 0020 and 0023 - 0024, Pg. 3 ¶ 0026 and 0030 - 0031, Pg. 4 ¶ 0034, Pg. 5 ¶ 0040 and 0044) and transmit instructions to cause a database to assign the predicted characteristic model including the predicted initial price to one or more data records associated with the target product. (Tibau-Puig et al., Figs. 1B & 3B - 7, Pg. 2 ¶ 0019 - 0021 and 0023 - 0024, Pg. 3 ¶ 0026, 0028 and 0030 - 0031, Pg. 4 ¶ 0034 and 0038, Pg. 5 ¶ 0040, Pg. 5 ¶ 0043 - Pg. 6 ¶ 0048, Pg. 6 ¶ 0052, Pg. 7 ¶ 0055 - 0057) Tibau-Puig et al. fail to disclose explicitly a similarity score; wherein images that are more similar to the target image receive a greater similarity score than a less similar image; identifying a similarity score above a threshold similarity score; weighting, for each similar product image in the set of similar product images, the historical event data associated with the respective similar product image based on the similarity score associated with the respective similar product image; and wherein a greater similarity score causes the historical event data associated with the respective similar product image to have a greater influence on a predicted characteristic model than historical event data associated with a different similar product image having a lower similarity score. Pertaining to analogous art, Ekambaram et al. disclose weighting, for each similar product image in the set of similar product images, the historical event data associated with the respective similar product image based on the distance score associated with the respective similar product image; (Ekambaram et al., Pg. 3112 § 3.1 - § 3.1.2 [“The estimator
ℇ
a produces a sales time-series estimate by performing weighted aggregation on time-series of k nearest neighbors”]) wherein a lower distance score (greater similarity score) causes the historical event data associated with the respective similar product image to have a greater influence on a predicted characteristic model than historical event data associated with a different similar product image having a greater distance score (lower similarity score); (Ekambaram et al., Pg. 3112 § 3.1 - § 3.1.2 [“The estimator
ℇ
a produces a sales time-series estimate by performing weighted aggregation on time-series of k nearest neighbors”, “represent the product features as a vector embedding. The vector embedding defines a proper distance metric. It is expected that similar products are close to each other compared to dissimilar products in this embedding space”, “X𝑝, and
Y
𝑝 represents the image/unstructured data and sales time-series attributed to product p. 𝑑 (Φ(X𝑝𝑖), Φ(X𝑝𝑗)) is the distance metric between product p𝑖, and p𝑗 in the embedding space” and “similar to attribute-based KNN, a KNN estimator ƐI is defined as
PNG
media_image1.png
60
420
media_image1.png
Greyscale
”. The Examiner asserts that one of ordinary skill in the art would understand that lower distances correspond to greater degrees of similarity, i.e., a lower distance score corresponds to a greater similarity score and vice-a-versa.]) and generating, by the machine learning model, the predicted characteristic model for the target product represented in the target product image based at least on the weighted historical event data associated with the set of similar product images, (Ekambaram et al., Pg. 3112 § 3.1 - § 3.1.2) the predicted characteristic model including a predicted initial price for the target product. (Ekambaram et al., Pg. 3111 § 1.1 ¶ 1, Pg. 3112 § 3.1, Pg. 3113 § 3.2.1 ¶ 1 - 2, Pg. 3115 § 4 ¶ 1) Ekambaram et al. fail to disclose expressly a similarity score; wherein images that are more similar to the target image receive a greater similarity score than a less similar image; and identifying a similarity score above a threshold similarity score. Pertaining to analogous art, Ikeda et al. disclose generating a similarity score between the target product image and one or more known product images from the group of known product images based at least on the comparison; (Ikeda et al., Figs. 2 & 8, Pg. 3 ¶ 0040 - 0043, Pg. 4 ¶ 0046 - 0048) wherein the known product images that are more similar to the target product image receive a greater similarity score than a less similar known product image; (Ikeda et al., Pg. 4 ¶ 0046 - 0048, Pg. 5 ¶ 0058 - 0059 [“assuming that the reciprocal of the distance is the similarity degree, the smaller the distance, the larger the similarity degree”]) and identifying a similarity score above a threshold similarity score. (Ikeda et al., Pg. 4 ¶ 0046 - 0048, Pg. 5 ¶ 0058 - 0059 [“determination section 54 compares the similarity degree calculated by the similarity degree calculation section 52 with a predetermined threshold value. If the similarity degree is larger than the threshold value, the determination section 54 determines that the object being recognized is identical with the reference object.”]) Craparotta et al. and Tibau-Puig et al. are combinable because they are both directed towards image processing systems that analyze images of products to estimate prices and/or demand for the products. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Craparotta et al. with the teachings of Tibau-Puig et al. These modifications would have been prompted in order to enhance the base device of Craparotta et al. with the well-known and applicable techniques Tibau-Puig et al. applied to a comparable device. Utilizing a computing system comprising at least one processor connected to at least one memory and a non-transitory computer readable medium including instructions stored thereon that, when executed by at least one processor, cause the at least one processor to perform functions, as taught by Tibau-Puig et al., to implement operations of the base device of Craparotta et al. would enhance the base device of Craparotta et al. by ensuring that its operations are carried out accurately and efficiently at high-computational speed on computer architecture, by facilitating widespread distribution of the base device of Craparotta et al. to millions of potential end-users with access to a computer and by simplifying the process of making revisions, modifications and/or updates to the operations of the base device of Craparotta et al. Additionally, causing a database to assign the predicted characteristic model including the predicted initial price to one or more data records associated with the target product, as taught by Tibau-Puig et al., would enhance the base device of Craparotta et al. by allowing for predicted characteristic models of target products to be preserved and readily available for future reference, analysis, and/or retrieval so as to provide end-users with the ability to subsequently utilize the predicted characteristic models of target products for any of a variety reasons and/or applications. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that a computing system comprising at least one processor connected to at least one memory and a non-transitory computer readable medium including instructions stored thereon that, when executed by at least one processor, cause the at least one processor to perform functions would be utilized to implement the base device of Craparotta et al. so as to ensure that its operations are carried out accurately and efficiently at high-computational speed on computer architecture, to facilitate widespread distribution of the base device of Craparotta et al. to millions of potential end-users with access to a computer and to simplify the process of making revisions, modifications and/or updates to the operations of the base device of Craparotta et al. and in that a database would assign the predicted characteristic model to one or more data records associated with the target product in order to allow for the predicted characteristic model of the target product to be preserved and readily available for future reference, analysis, and/or retrieval so as to provide end-users with the ability to subsequently utilize the predicted characteristic model of the target product for any of a variety reasons. In addition, Craparotta et al. in view of Tibau-Puig et al. and Ekambaram et al. are combinable because they are all directed towards image processing systems that analyze images of products to estimate prices and/or demand for the products. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. with the teachings of Ekambaram et al. This modification would have been prompted in order to enhance the combined base device of Craparotta et al. in view of Tibau-Puig et al. with the well-known and applicable technique Ekambaram et al. applied to a comparable device. Weighting the historical event data for each similar object image with the distance score associated with each of the similar object images when generating the predicted characteristic model for the target object, as taught by Ekambaram et al., would enhance the combined base device by helping improve its ability to accurately and reliably predict the sales profile of an imaged product since sales profiles for products used in predicting the sales profile of the imaged product would be weighted so that sales profiles for products that are the most similar to the imaged product have more of an influence on the predicted sales profile than sales profiles of less similar products thereby allowing for products which are more comparable to the imaged product to have more of an impact on the predicted sales profile of the imaged product. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that historical event data for each similar object image would be weighted with the distance score associated with each of the similar object images when generating the predicted characteristic model for the target object so as to allow for objects which are more comparable to the target object than others to have more of an influence on the predicted characteristic model for the target object than less comparable objects in order to help improve the ability of the combined base device to accurately and reliably generate the predicted characteristic model for the target object. Additionally, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. and Ikeda et al. are combinable because they are all directed towards image processing systems that classify images of objects and, similar to Craparotta et al. and Ekambaram et al., Ikeda et al. also identify images that are similar to an input image. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. with the teachings of Ikeda et al. This modification would have been prompted in order to substitute the distance score of the combined base device for the similarity degree of Ikeda et al. The similarity degree of Ikeda et al. could be substituted in place of the distance score of the combined base device utilizing well-known techniques in the art and would likely yield predictable results, in that in the combination the similarity degree of Ikeda et al. would be utilized to define the similarity between images/products, identify similar images and weight the historical event data. Furthermore, modification would have been prompted by the teachings and suggestions of Ekambaram et al. that their vector embedding defines a proper distance metric, that it is expected that similar products are close to each other compared to dissimilar products in their embedding space and that the similarity between products can be defined based on the cosine distance between their image embeddings, see at least pages 3112 - 3113 section 3.1 and page 3116 section 4.2 paragraph 1 of Ekambaram et al. Moreover, this modification would have been prompted by the teachings and suggestions of Ikeda et al. that the reciprocal of a distance between images can be treated as a similarity degree between images, see at least page 4 paragraphs 0046 - 0048 of Ikeda et al. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that the similarity degree of Ikeda et al., the reciprocal of distance, would be used in place of the distance score of the combined base device and utilized by the combined base device to define the similarity between images/products, identify similar images and weight the historical event data. Therefore, it would have been obvious to combine Craparotta et al. with Tibau-Puig et al., Ekambaram et al. and Ikeda et al. to obtain the invention as specified in claim 15.
- With regards to claim 16, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. disclose the computing system of claim 15, wherein the instructions further include instructions that, when executed by at least the at least one processor, cause the at least one processor to: analyze, by the machine learning model, the digital pixel data of the target product image including object-based image analysis to group pixels to identify the target product in the target product image. (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph - Pg. 1539 § 2.2, Pg. 1539 Fig. 3, Pg. 1541 Left-Hand Column First-Full Paragraph - Fifth-Full Paragraph, Pg. 1541 Fig. 7, Pg. 1542 § 4 - § 4.1 ¶ 2, Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) In addition, analogous art Tibau-Puig et al. disclose instructions that, when executed by at least the at least one processor, cause the at least one processor to: analyze, by the machine learning model, the digital pixel data of the target product image including object-based image analysis to group pixels to identify the target product in the target product image. (Tibau-Puig et al., Pg. 1 ¶ 0016 - Pg. 2 ¶ 0017, Pg. 2 ¶ 0021 - 0024, Pg. 5 ¶ 0039 and 0042 - 0043)
- With regards to claim 17, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. disclose the computing system of claim 15, wherein the instructions further include instructions that, when executed by at least the at least one processor, cause the at least one processor to: generate an electronic message with the predicted characteristic for the target product. (Craparotta et al., Pg. 1538 Right-Hand Column First-Full Paragraph - Second-Full Paragraph, Pg. 1538 Fig. 2, Pg. 1540 § 3 - § 3.1, Pg. 1540 Fig. 5, Pg. 1541 § 3.1.3, Pg. 1542 § 4.1 - Pg. 1543 § 4.2, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) Craparotta et al. fail to disclose explicitly transmitting the electronic message to a remote computer. Pertaining to analogous art, Tibau-Puig et al. disclose instructions that, when executed by at least the at least one processor, cause the at least one processor to: generate an electronic message with the predicted characteristic for the target product; (Tibau-Puig et al., Abstract, Figs. 1A & 3B - 5, Pg. 2 ¶ 0019, Pg. 2 ¶ 0023 - Pg. 3 ¶ 0027, Pg. 4 ¶ 0032 - 0034 and 0038, Pg. 5 ¶ 0040 - 0042, Pg. 6 ¶ 0054) and transmit the electronic message to a remote computer. (Tibau-Puig et al., Figs. 1A, 1B, 3B & 4, Pg. 2 ¶ 0024 - Pg. 3 ¶ 0027, Pg. 4 ¶ 0032 - 0034 and 0038, Pg. 6 ¶ 0050 - 0051 and 0053 - 0054) It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with additional teachings of Tibau-Puig et al. This modification would have been prompted in order to enhance the combined base device of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the well-known and applicable technique Tibau-Puig et al. applied to a comparable device. Transmitting the electronic message to a remote computer, as taught by Tibau-Puig et al., would enhance the combined base device by enabling results obtained by the combined base device, the electronic message, to be easily shared between different users and/or locations and/or by allowing for end-users with devices having low computational processing power to offload most or all of the computationally intensive processing tasks to a remotely located and computationally powerful computing system so as to increase the overall operational speed of the combined base device. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that the electronic message would be transmitted to a remote computer so as to enable the electronic message to be easily shared between different users and/or locations and/or to allow for end-users with devices having low computational processing power to offload most or all of the computationally intensive processing tasks to a remotely located and computationally powerful computing system so as to increase the overall operational speed of the combined base device. Therefore, it would have been obvious to combine Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with additional teachings of Tibau-Puig et al. to obtain the invention as specified in claim 17.
Claims 6, 13 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Giuseppe Craparotta, Sébastien Thomassey, Amedeo Biolatti, "A Siamese Neural Network Application for Sales Forecasting of New Fashion Products Using Heterogeneous Data", International Journal of Computational Intelligence Systems, Vol. 12(2), Nov. 2019, pages 1537 - 1546, herein referred to as “Craparotta et al.”, in view of Tibau-Puig et al. U.S. Publication No. 2019/0294879 A1 in view of Vijay Ekambaram, Kushagra Manglik, Sumanta Mukherjee, Surya Shravan Kumar Sajja, Satyam Dwivedi, Vikas Raykar, "Attention based Multi-Modal New Product Sales Time-series Forecasting", Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, Aug. 2020, pages 3110 - 3118, herein referred to as “Ekambaram et al.”, in view of Ikeda et al. U.S. Publication No. 2005/0238209 A1 as applied to claims 1, 8 and 15 above, and further in view of Stefanescu et al. U.S. Publication No. 2003/0013951 A1.
- With regards to claim 6, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. disclose the method of claim 1, comprising: identifying the set of similar object images. (Craparotta et al., Pg. 1541 Left-Hand Column First-Full Paragraph - § 3.1.3, Pg. 1542 § 4.1 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5, Pg. 1545 Fig. 16) Craparotta et al. fail to disclose explicitly removing similar object images from the set of similar object images that do not contain an image orientation that is similar to an image orientation in the target object image. Pertaining to analogous art, Stefanescu et al. disclose removing similar object images from the set of similar object images that do not contain an image orientation that is similar to an image orientation in the target object image. (Stefanescu et al., Pg. 1 ¶ 0024, Pg. 5 ¶ 0053, Pg. 5 ¶ 0056 - Pg. 6 0058, Pg. 6 ¶ 0061 and 0064, Pg. 7 ¶ 0073 - Pg. 8 ¶ 0076, Pg. 11 ¶ 0099 [“Orientation and sequence filtering may remove images obtained, for example, on a different axis from the query image, or using a different imaging modality.”]) Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. and Stefanescu et al. are combinable because they are all directed towards image processing systems that classify images of objects and, similar to Craparotta et al., Ekambaram et al. and Ikeda et al., Stefanescu et al. also identify images that are similar to an input image. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the teachings of Stefanescu et al. This modification would have been prompted in order to enhance the combined base device of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the well-known and applicable technique Stefanescu et al. applied to a similar device. Removing similar object images that do not contain an image orientation that is similar to an image orientation in the target object image, as taught by Stefanescu et al., would enhance the combined base device by improving its ability to quickly, efficiently and reliably identify object images, and thus objects, that are the most similar and comparable to the target object image since object images that do not contain an image orientation similar to an image orientation of the target object image would be prevented from being included in the set of similar object images thereby eliminating unnecessary object images from undergoing further analysis and reducing the overall number of computations required to implement the combined base device. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that similar object images that do not contain an image orientation that is similar to an image orientation in the target object image would be removed and prevented from being included in the set of similar object images so as to eliminate unnecessary object images from undergoing further analysis and reduce the overall number of computations required to implement the combined base device thereby improving the overall efficiency and operational speed of the combined base device. Therefore, it would have been obvious to combine Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with Stefanescu et al. to obtain the invention as specified in claim 6.
- With regards to claim 13, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. disclose the non-transitory computer-readable medium of claim 8. Craparotta et al. fail to disclose explicitly instructions that, when executed by at least the processor, cause the processor to: remove similar product images from the set of similar product images that do not contain an image orientation that is similar to an image orientation in the target product image. Pertaining to analogous art, Stefanescu et al. disclose instructions that, when executed by at least the processor, cause the processor to: remove similar product images from the set of similar product images that do not contain an image orientation that is similar to an image orientation in the target product image. (Stefanescu et al., Pg. 1 ¶ 0024, Pg. 5 ¶ 0053, Pg. 5 ¶ 0056 - Pg. 6 0058, Pg. 6 ¶ 0061 and 0064, Pg. 7 ¶ 0073 - Pg. 8 ¶ 0076, Pg. 11 ¶ 0099 [“Orientation and sequence filtering may remove images obtained, for example, on a different axis from the query image, or using a different imaging modality.”]) Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. and Stefanescu et al. are combinable because they are all directed towards image processing systems that classify images of objects and, similar to Craparotta et al., Ekambaram et al. and Ikeda et al., Stefanescu et al. also identify images that are similar to an input image. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the teachings of Stefanescu et al. This modification would have been prompted in order to enhance the combined base device of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the well-known and applicable technique Stefanescu et al. applied to a similar device. Removing similar product images that do not contain an image orientation that is similar to an image orientation in the target product image, as taught by Stefanescu et al., would enhance the combined base device by improving its ability to quickly, efficiently and reliably identify product images, and thus product, that are the most similar and comparable to the target product image since product images that do not contain an image orientation similar to an image orientation of the target product image would be prevented from being included in the set of similar product images thereby eliminating unnecessary product images from undergoing further analysis and reducing the overall number of computations required to implement the combined base device. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that similar product images that do not contain an image orientation that is similar to an image orientation in the target product image would be removed and prevented from being included in the set of similar product images so as to eliminate unnecessary product images from undergoing further analysis and reduce the overall number of computations required to implement the combined base device thereby improving the overall efficiency and operational speed of the combined base device. Therefore, it would have been obvious to combine Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with Stefanescu et al. to obtain the invention as specified in claim 13.
- With regards to claim 19, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. disclose the computing system of claim 15. Craparotta et al. fail to disclose explicitly instructions that, when executed by at least the at least one processor, cause the at least one processor to: remove similar product images from the set of similar product images that do not contain an image orientation that is similar to an image orientation in the target product image. Pertaining to analogous art, Stefanescu et al. disclose instructions that, when executed by at least the at least one processor, cause the at least one processor to: remove similar product images from the set of similar product images that do not contain an image orientation that is similar to an image orientation in the target product image. (Stefanescu et al., Pg. 1 ¶ 0024, Pg. 5 ¶ 0053, Pg. 5 ¶ 0056 - Pg. 6 0058, Pg. 6 ¶ 0061 and 0064, Pg. 7 ¶ 0073 - Pg. 8 ¶ 0076, Pg. 11 ¶ 0099 [“Orientation and sequence filtering may remove images obtained, for example, on a different axis from the query image, or using a different imaging modality.”]) Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. and Stefanescu et al. are combinable because they are all directed towards image processing systems that classify images of objects and, similar to Craparotta et al., Ekambaram et al. and Ikeda et al., Stefanescu et al. also identify images that are similar to an input image. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the teachings of Stefanescu et al. This modification would have been prompted in order to enhance the combined base device of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the well-known and applicable technique Stefanescu et al. applied to a similar device. Removing similar product images that do not contain an image orientation that is similar to an image orientation in the target product image, as taught by Stefanescu et al., would enhance the combined base device by improving its ability to quickly, efficiently and reliably identify product images, and thus product, that are the most similar and comparable to the target product image since product images that do not contain an image orientation similar to an image orientation of the target product image would be prevented from being included in the set of similar product images thereby eliminating unnecessary product images from undergoing further analysis and reducing the overall number of computations required to implement the combined base device. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that similar product images that do not contain an image orientation that is similar to an image orientation in the target product image would be removed and prevented from being included in the set of similar product images so as to eliminate unnecessary product images from undergoing further analysis and reduce the overall number of computations required to implement the combined base device thereby improving the overall efficiency and operational speed of the combined base device. Therefore, it would have been obvious to combine Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with Stefanescu et al. to obtain the invention as specified in claim 19.
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Giuseppe Craparotta, Sébastien Thomassey, Amedeo Biolatti, "A Siamese Neural Network Application for Sales Forecasting of New Fashion Products Using Heterogeneous Data", International Journal of Computational Intelligence Systems, Vol. 12(2), Nov. 2019, pages 1537 - 1546, herein referred to as “Craparotta et al.”, in view of Tibau-Puig et al. U.S. Publication No. 2019/0294879 A1 in view of Vijay Ekambaram, Kushagra Manglik, Sumanta Mukherjee, Surya Shravan Kumar Sajja, Satyam Dwivedi, Vikas Raykar, "Attention based Multi-Modal New Product Sales Time-series Forecasting", Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, Aug. 2020, pages 3110 - 3118, herein referred to as “Ekambaram et al.”, in view of Ikeda et al. U.S. Publication No. 2005/0238209 A1 as applied to claim 1 above, and further in view of Gupta et al. U.S. Publication No. 2022/0398528 A1.
- With regards to claim 7, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. disclose the method of claim 1. Craparotta et al. fail to disclose explicitly causing a robotic mechanism to retrieve quantities of the target object from a storage location based at least upon the predicted characteristic model. Pertaining to analogous art, Gupta et al. disclose causing a robotic mechanism to retrieve quantities of the target object from a storage location based at least upon the predicted characteristic model. (Gupta et al., Abstract, Fig. 7B, Pg. 3 ¶ 0033 - Pg. 4 ¶ 0037, Pg. 5 ¶ 0048 - Pg. 6 ¶ 0049, Pg. 9 ¶ 0077 - 0078, Pg. 13 ¶ 0102 and 0105) Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. and Gupta et al. are combinable because, similar to Craparotta et al., Tibau-Puig et al. and Ekambaram et al., Gupta et al. is also directed towards predicting levels of demand for products. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the teachings of Gupta et al. This modification would have been prompted in order to enhance the combined base device of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the well-known and applicable technique Gupta et al. applied to a similar device. Causing a robotic mechanism to retrieve quantities of the target object from a storage location based at least upon the predicted characteristic model, as taught by Gupta et al., would enhance the combined base device by allowing for it to be utilized in an increased number of practical applications, such as inventory management applications, thereby improving the overall appeal and usefulness of the combined base device to potential end-users. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that a robotic mechanism would retrieve quantities of the target object from a storage location based at least upon the predicted characteristic model so as to allow for the combined base device to be utilized in an increased number of practical applications, such as inventory management applications, in order to improve the overall appeal and usefulness of the combined base device to potential end-users. Therefore, it would have been obvious to combine Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with Gupta et al. to obtain the invention as specified in claim 7.
Claims 11 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Giuseppe Craparotta, Sébastien Thomassey, Amedeo Biolatti, "A Siamese Neural Network Application for Sales Forecasting of New Fashion Products Using Heterogeneous Data", International Journal of Computational Intelligence Systems, Vol. 12(2), Nov. 2019, pages 1537 - 1546, herein referred to as “Craparotta et al.”, in view of Tibau-Puig et al. U.S. Publication No. 2019/0294879 A1 in view of Vijay Ekambaram, Kushagra Manglik, Sumanta Mukherjee, Surya Shravan Kumar Sajja, Satyam Dwivedi, Vikas Raykar, "Attention based Multi-Modal New Product Sales Time-series Forecasting", Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, Aug. 2020, pages 3110 - 3118, herein referred to as “Ekambaram et al.”, in view of Ikeda et al. U.S. Publication No. 2005/0238209 A1 as applied to claims 8 and 15 above, and further in view of Davis et al. U.S. Publication No. 2022/0198496 A1.
- With regards to claim 11, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. disclose the non-transitory computer-readable medium of claim 8, further comprising instructions that, when executed by at least the processor, cause the processor to: adjust the predicted characteristic by at least the similarity score of a similar product image of the set of similar product images. (Craparotta et al., Pg. 1537 Abstract, Pg. 1538 Left-Hand Column Second-Full Paragraph - Right-Hand Column First-Full Paragraph, Pg. 1540 § 3 - § 3.1.1, Pg. 1540 Figs. 5 & 6, Pg. 1541 § 3.1.3, Pg. 1542 § 4 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5) Craparotta et al. fail to disclose explicitly negatively adjusting the predicted characteristic by at least the similarity score that is associated with a selected brand of a known product. Pertaining to analogous art, Davis et al. disclose instructions that, when executed by at least the processor, cause the processor to: negatively adjust the predicted characteristic by at least the similarity score of a similar product image of the set of similar product images that is associated with a selected brand of a known product. (Davis et al., Pg. 2 ¶ 0018, Pg. 5 ¶ 0054 and 0056 - 0059, Pg. 6 ¶ 0068 - 0070) Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. and Davis et al. are combinable because, similar to Craparotta et al., Tibau-Puig et al. and Ekambaram et al., Davis et al. is also directed towards estimating information for an input product by identifying products similar to the input product and analyzing historical data of the identified similar products. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the teachings of Davis et al. This modification would have been prompted in order to enhance the combined base device of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the well-known and applicable technique Davis et al. applied to a similar device. Negatively adjusting the predicted characteristic by at least the similarity score that is associated with a selected brand of a known product, as taught by Davis et al., would enhance the combined base device by allowing for it to take into account and adjust for differences between a brand of an imaged product and brands of identified similar products when generating the predicted characteristic, sales profile, for the imaged product thereby helping ensure that the sales profile it predicts for the imaged product is as accurate and reliable as possible. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that the predicted characteristic would be negatively adjusted by at least the similarity score that is associated with a selected brand of a known product so as to allow for the combined base device to take into account and adjust for differences between a brand of an imaged product and brands of identified similar products when generating the predicted characteristic, sales profile, for the imaged product. Therefore, it would have been obvious to combine Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with Davis et al. to obtain the invention as specified in claim 11.
- With regards to claim 20, Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. disclose the computing system of claim 15, wherein the instructions further include instructions that, when executed the at least one processor, cause the at least one processor to: adjust the predicted characteristic by at least the similarity score of a similar product image of the set of similar product images. (Craparotta et al., Pg. 1537 Abstract, Pg. 1538 Left-Hand Column Second-Full Paragraph - Right-Hand Column First-Full Paragraph, Pg. 1540 § 3 - § 3.1.1, Pg. 1540 Figs. 5 & 6, Pg. 1541 § 3.1.3, Pg. 1542 § 4 - Pg. 1543 § 4.2 ¶ 1, Pg. 1543 Fig. 11, Pgs. 1544 - 1545 § 5) Craparotta et al. fail to disclose explicitly negatively adjusting the predicted characteristic by at least the similarity score that is associated with a selected brand of a known product. Pertaining to analogous art, Davis et al. disclose instructions that, when executed by the at least one processor, cause the at least one processor to: negatively adjust the predicted characteristic by at least the similarity score of a similar product image of the set of similar product images that is associated with a selected brand of a known product. (Davis et al., Pg. 2 ¶ 0018, Pg. 5 ¶ 0054 and 0056 - 0059, Pg. 6 ¶ 0068 - 0070) Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. and Davis et al. are combinable because, similar to Craparotta et al., Tibau-Puig et al. and Ekambaram et al., Davis et al. is also directed towards estimating information for an input product by identifying products similar to the input product and analyzing historical data of the identified similar products. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined teachings of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the teachings of Davis et al. This modification would have been prompted in order to enhance the combined base device of Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with the well-known and applicable technique Davis et al. applied to a similar device. Negatively adjusting the predicted characteristic by at least the similarity score that is associated with a selected brand of a known product, as taught by Davis et al., would enhance the combined base device by allowing for it to take into account and adjust for differences between a brand of an imaged product and brands of identified similar products when generating the predicted characteristic, sales profile, for the imaged product thereby helping ensure that the sales profile it predicts for the imaged product is as accurate and reliable as possible. This combination could be completed according to well-known techniques in the art and would likely yield predictable results, in that the predicted characteristic would be negatively adjusted by at least the similarity score that is associated with a selected brand of a known product so as to allow for the combined base device to take into account and adjust for differences between a brand of an imaged product and brands of identified similar products when generating the predicted characteristic, sales profile, for the imaged product. Therefore, it would have been obvious to combine Craparotta et al. in view of Tibau-Puig et al. in view of Ekambaram et al. in view of Ikeda et al. with Davis et al. to obtain the invention as specified in claim 20.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
a. Banipal et al. U.S. Publication No. 2022/0188852 A1; which is directed towards a method and system for determining optimal prices for products, wherein a machine learning model predicts an optimal price for a product based on visual textual data of the product.
b. Perez-Rovira et al. U.S. Publication No. 2023/0222826 A1; which is directed towards a method and system that determines the similarity in appearance between images, wherein a distance between low dimensionality vectors of two images is calculated and the similarity in appearance between the two images is determined by using the inverse of the distance.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ERIC RUSH whose telephone number is (571) 270-3017. The examiner can normally be reached 9am - 5pm Monday - Friday.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Andrew Bee can be reached at (571) 270 - 5183. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/ERIC RUSH/Primary Examiner, Art Unit 2677
Prosecution Insights