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 .
Claims Status
Claims 14-20 have been cancelled.
Claims 25-27 are newly added.
Claims 1-13 and 21-27 are pending and stand rejected.
Examiner Comment: Proposed Amendment
The Examiner draws Applicant’s attention to the proposed amendment under the heading Appendix – Proposed Amendment on page 23 of this correspondence (below).
Response to Arguments
I. Applicant’s arguments made with respect to the rejection under 35 USC 101 have been fully considered but are not persuasive.
Applicant’s arguments with respect to Step 2A (Prong One) are acknowledged but the Examiner disagrees. Step 2A is a two-prong inquiry, in which examiners determine in Prong One whether a claim recites a judicial exception, and if so, then determine in Prong Two if the recited judicial exception is integrated into a practical application of that exception.
In Prong One examiners evaluate whether the claim recites a judicial exception, i.e. whether a law of nature, natural phenomenon, or abstract idea is set forth or described in the claim. The Examiner maintains that the claims clearly set forth or describe an abstract idea, and thus “recite” an abstract idea (Prong One = Yes). The Examiner reiterates that the claimed invention sets forth or describes perform an item search in a marketplace using an image. This falls squarely within ‘certain methods of organizing human activity’ because the claims clearly set forth or describe concepts relating to the economy and commerce. Although the exemplary decisions may differ in their specific claim language or in the concept articulated by the courts, Applicant is reminded that the provided examples are simply that - examples of the concepts the courts have consistently identified as abstract, and are not limiting. Both the MPEP and the courts have declined to define abstract ideas, other than by example. Both the office and the courts have provided consistent guidance instructing examiners to refer to the body of case law precedent in order to identify abstract ideas by way of comparison to concepts already found to be abstract. This was expressly provided in prior guidance, and has been reiterated in MPEP 2106.04(a):
To facilitate examination, the Office has set forth an approach to identifying abstract ideas that distills the relevant case law into enumerated groupings of abstract ideas. The enumerated groupings are firmly rooted in Supreme Court precedent as well as Federal Circuit decisions interpreting that precedent, as is explained in MPEP § 2106.04(a)(2). This approach represents a shift from the former case-comparison approach that required examiners to rely on individual judicial cases when determining whether a claim recites an abstract idea. By grouping the abstract ideas, the examiners’ focus has been shifted from relying on individual cases to generally applying the wide body of case law spanning all technologies and claim types.
Applicant’s allegations run contrary to the Examination Guidance in the MPEP and contrary to the manner in which the courts have assessed claim eligibility. While distinctions may exist between certain decisions the claimed invention, the claimed invention clearly sets forth limitations that fall within the bounds of commercial processes such as marketing or sales activities, which are enumerated abstract concepts, and thus fall under certain methods of organizing human activity. The Examiner previous rejection, as well as the updated rejection below, clearly identifies the limitations which “recite” the abstract idea(s). Any additional elements are then analyzed under Prong Two, which is addressed further below.
With respect to Prong Two, the Examiner again disagrees. That is, the Examiner fundamentally disagrees that analyzing an image to ascertain product features and performing a search should be considered an improvement to the functioning of the computer itself or another technology or technical field. Most notably, “enabling more accurate and intent-driven search results” is not an improvement to a technical field, but instead an improvement to the abstract commercial process itself. The mere use of machine learning in automating the image analysis does not move this from a abstract idea improvement into a technological one. This is because the claimed invention provides no restriction on how the underlying operations of the machine learning are performed, instead relying solely on highly-generalized machine learning techniques to achieve the abstract analysis more efficiently or accurately. The recitation of claim limitations that attempt to cover any solution to an identified problem with no restriction on how the result is accomplished and no description of the mechanism for accomplishing the result, does not integrate a judicial exception into a practical application or provide significantly more because this type of recitation is equivalent to the words “apply it”.
As in previous actions, the Examiner again emphasizes that the disclosure must provide sufficient details such that one of ordinary skill in the art would recognize the claimed invention as providing an improvement. The Examiner also emphasizes that the Federal Circuit has stressed that "merely using a computer to perform more efficiently what could otherwise be accomplished manually does not confer patent-eligibility." See buySAFE, Inc. v. Google, Inc., 964 F. Supp. 2d 331, 336 (D. Del. 2013) (citing Bancorp Servs., L.L.C. v. Sun Life Assur. Co. of Can., 687 F.3d 1266, 1279 (Fed. Cir. 2012)), aff'd, 765 F.3d 1350 (Fed. Cir. 2014). Similarly, “claiming the improved speed or efficiency inherent with applying the abstract idea on a computer” does not integrate a judicial exception into a practical application or provide an inventive concept (Intellectual Ventures I LLC v. Capital One Bank (USA), 792 F.3d 1363, 1367, 115 USPQ2d 1636, 1639 (Fed. Cir. 2015)).
Similar analysis as applied under Prong Two is also applied under Step 2B. Taken individually or as a whole the additional elements of claim 1 do not provide an inventive concept (i.e. they do not amount to “significantly more” than the exception itself). As discussed above with respect to the integration of the abstract idea into a practical application, the additional elements used to perform the claimed process amount to no more than the mere instructions to apply the exception using a generic computer, no more than a general link to a technological environment, and operate using well-understood, routine and conventional computer operations, such as receiving or transmitting data over a network, performing repetitive calculations, and presenting offers (see MPEP 2106.05(d)(II)).. Even considered as an ordered combination (as a whole), the additional elements of claim 1 do not add anything further than when they are considered individually.
Regarding new claim 27, the Examiner notes that claim 27 is similar to claim 6. only adds noise to the data without performing any subsequent training, as does claim 27. As written, it is not inherent that the model is iteratively or even actively trained, and neither of these claims reflect any form of active training by adding/injecting noise (e.g., after an initial training, or during an active training phase). Moreover, paragraph 0087, however, describes using known types of noise (e.g., Gaussian Noise, Poisson Noise) that achieve known improvements as an ancillary part of performing the abstract idea.
Even assuming arguendo the paragraphs do demonstrate an inventive improvement by Applicant (which the Examiner does not acquiesce), the claims as written do not achieve this improvement because they do not positively recite training using the noise data (which is required to achieve the improvement), and these elements simply extend the mere use of a [previously] trained machine learning model (as discussed with respect to claim 1). Ultimately, neither claim 27 nor claim 6 achieve an improvement to the image recognition because the limitations are recited passively in the past tense, and merely uses an already-trained model/CNN in automating the analysis. Moreover, PTO form 892-U (previously cited in the correspondence mailed 2/12/2026) demonstrates the conventionality of deliberately introducing noise to help improve computer vision models (i.e., adding noise), including various known types of noise that may be added (e.g., Gaussian, Localvar, Poisson, Salt, Pepper, s&p, speckle) (see: Paragraphs 3-4, Images under “Implementing Noise”). Here again, the Examiner re-emphasizes that the addition of these features represents a post-hoc addition of known features for their known benefits in support of the abstract idea, and does not demonstrate that Applicant’s invention achieves an improvement in a technical field. Accordingly, the rejection under 35 USC 101 has been maintained below.
II. Applicant’s arguments made with respect to the rejections under 35 USC 103 have ben fully considered but are not persuasive. Applicant argues that the combination of Ramesh and Badjatiya presenting the one or more search terms in a user interface configured to receive an edit to the one or more search terms, including an edit to the type of object. The Examiner respectfully disagrees.
Ramesh specifically discloses determining product categories from the image data and providing the category as a keyword (see: 0025, 0032), such as in an alternative keywork recommendation portion of a user interface (see: Fig. 5a, 0027). The alternative keyword searches are presented for selection by the user (see: Fig. 5a (510, 518-520), 0029-0030, Fig. 7 (710)) and enable refining of the initial search (see: 0030, 0015, Fig. 7 (714-718)). This refinement represents an edit, which occurs from within the user interface. Moreover, because category (i.e., type) of product may be included in this portion of the interface enabling refinement, Ramesh also discloses refinements including an edit to the type of object. Accordingly, Ramesh discloses presenting the one or more search terms in a user interface configured to receive an edit to the one or more search terms, including an edit to the type of object
Though not rendered obvious over Ramesh and Badjatiya alone, new claims 25-26 have necessitated new grounds of rejection. Applicant’s amendment necessitated these new grounds.
Regarding new claim 27, the Examiner notes Badjatiya, teaches the machine learning model to include a convolutional neural network (see: 0088, 0091-0092, 0101). The Examiner also reiterates that the ML model is trained as discussed in the previous rejection (see again: 0022, 0038, 0040, 0042, 0067, 0070, Fig. 1, Fig. 9 (960)).
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-13 and 21-27 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (abstract idea) without significantly more.
Regarding claims 1-13 and 21-27, under Step 2A claims 1-13 and 21-27 recite a judicial exception (abstract idea) that is not integrated into a practical application and does not provide significantly more.
Under Step 2A (prong 1), and taking claim 1 as representative, claim 1 recites
a method comprising:
receiving a search query for items listed on a marketplace, the search query including an image;
analyzing the image to determine characteristics of an object in the image;
automatically generating one or more search terms, including a type of the object, based on the characteristics of the object in the image;
receive an edit to the one or more search terms, including the type of object;
transmitting the one or more search terms to the marketplace to locate items matching the one or more search terms; and
displaying visual indications of located items matching the one or more search terms in the marketplace.
These limitations recite ‘certain methods of organizing human activity’, such as by performing commercial interactions (see: MPEP 2106.04(a)(2)(II)). This is because claim 1 recites searching for items in a marketplace using an image. This represents the performance of marketing or sales activities or behaviors, which is a commercial interaction and falls under organizing human activity.
Accordingly, under step 2A (prong 1) claim 1 recites an abstract idea because claim 1 recites limitations that fall within the “Certain methods of organizing human activity” grouping of abstract ideas.
Under Step 2A (prong 2), the abstract idea is not integrated into a practical application. The Examiner acknowledges that representative claim 1 does recite additional elements, including:
a computer-implemented method,
an online marketplace,
using one or more machine learning models including a convolutional neural network trained to extract visual features representing the characteristics of the object, and,
presenting a user interface,
transmitting search terms over a network and to one or more servers.
Although reciting these additional elements, taken alone or in combination these elements are not sufficient to integrate the abstract idea into a practical application. This is because the additional elements of claim 1 are recited at a high level of generality (i.e. as generic computing hardware) such that they amount to nothing more than the mere instructions to implement or apply the abstract idea on generic computing hardware (or, merely uses a computer as a tool to perform an abstract idea). This is true even with respect to using one or more machine learning models including a convolutional neural network trained to extract visual features representing the characteristics of the object, which merely leverages this additional element for perform the abstract idea of analyzing images and determining characteristics. That is, each of the additional elements (including the use of the machine learning model), taken alone or in combination, represents nothing more than applying the abstract idea using generic computing components such that they are merely uses a computer as a tool to perform an abstract idea.
Further, the additional elements (e.g., “online” marketplace) do no more than generally link the use of a judicial exception to a particular technological environment or field of use (such as the Internet or computing networks).
Lastly, the additional elements are insufficient to integrate the abstract idea into a practical application because the claim fails to (i) reflect an improvement in the functioning of a computer, or an improvement to other technology or technical field, (ii) implement the judicial exception with, or use the judicial exception in conjunction with, a particular machine or manufacture that is integral to the claim, (iii) effect a transformation or reduction of a particular article to a different state or thing, or (iv) applies or uses the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment.
In view of the above, under Step 2A (prong 2), claim 1 does not integrate the recited exception into a practical application.
Under Step 2B, examiners should evaluate additional elements individually and in combination to determine whether they provide an inventive concept (i.e., whether the additional elements amount to significantly more than the exception itself). In this case, the claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Returning to representative claim 1, taken individually or as a whole the additional elements of claim 1 do not provide an inventive concept (i.e. they do not amount to “significantly more” than the exception itself). As discussed above with respect to the integration of the abstract idea into a practical application, the additional elements used to perform the claimed process amount to no more than the mere instructions to apply the exception using a generic computer, or no more than a general link to a technological environment.
Additionally, the additional elements operate using well-understood, routine and conventional computer operations, such as receiving or transmitting data over a network, performing repetitive calculations, and presenting offers (see MPEP 2106.05(d)(II)). Even considered as an ordered combination (as a whole), the additional elements of claim 1 do not add anything further than when they are considered individually.
In view of the above, representative claim 1 does not provide an inventive concept (“significantly more”) under Step 2B, and is therefore ineligible for patenting.
Regarding dependent claims 2-9 and 25-27, dependent claims 2-9 and 25-27 recite more complexities descriptive of the abstract idea itself, and at least inherit the abstract idea of claim 1. As such, claims 2-9 and 25-27 are understood to recite an abstract idea under step 2A (prong 1) for at least similar reasons as discussed above.
Under prong 2 of step 2A, the additional elements of dependent claims 2-9 and 25-27 also do not integrate the abstract idea into a practical application, considered both individually or as a whole. More specifically, claims 2-9 rely upon at least elements as recited in claim 1. Further additional elements (e.g., checkboxes of claim 25, dropdown menu of claim 26) are also recited only at a high level of generality (i.e. as generic computing hardware) such that they amount to nothing more than the mere instructions to implement or apply the abstract idea on generic computing hardware (or, merely uses a computer as a tool to perform an abstract idea). Further, the additional elements do no more than generally link the use of a judicial exception to a particular technological environment or field of use (such as the Internet or computing networks).
With specific reference to claims 5-6 and 27, which recite aspects related to machine learning, the Examiner maintains that the claims do not integrate the abstract idea into a practical application. Claim 5 recites that the machine learning model is trained using training data, while claim 6 recites “adding noise”. Claim 6 only adds noise to the data without performing any subsequent training, as does claim 27. As written, it is not inherent that the model is iteratively or even actively trained, and none of these claims reflect any form of active training by adding/injecting noise (e.g., after an initial training, or during an active training phase). Moreover, paragraph 0087, however, describes using known types of noise (e.g., Gaussian Noise, Poisson Noise) that achieve known improvements as an ancillary part of performing the abstract idea.
Even assuming arguendo the paragraphs do demonstrate an inventive improvement by Applicant (which the Examiner does not acquiesce), the claims as written do not achieve this improvement because they do not positively recite training using the noise data (which is required to achieve the improvement), and these elements simply extend the mere use of a [previously] trained machine learning model (as discussed with respect to claim 1). Accordingly, claims 2-9 and 25-27 do not integrate the recited exception into a practical application.
Lastly, under step 2B, claims 2-9 and 25-27 also fail to result in “significantly more” than the abstract idea under step 2B. This is again because the claims merely apply the exception on generic computing hardware and/or generally link the exception to a technological environment. Even when viewed as an ordered combination (as a whole), the additional elements of the dependent claims do not add anything further than when they are considered individually.
In view of the above, claims 2-9 and 25-27 do not provide an inventive concept (“significantly more”) under Step 2B, and are therefore ineligible for patenting.
Regarding claims 10-13 (system) and claims 21-24 (non-transitory CRSM), claims 10-13 and 21-24 recite at least substantially similar concepts and elements as recited in claims 1-9 and 25-27 such that similar analysis of the claims would be readily apparent to one of ordinary skill in the art. Further additional elements such as memory, storage media, instructions, et al. similarly represent mere use of generic computing components to implement the abstract idea. As such, claims 10-16 and 21-24 are rejected under at least similar rationale.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1-2, 4, 7, 9-11, 21-22, 24 and 27 is/are rejected under 35 U.S.C. 103 as being unpatentable over Ramesh (US 2020/0311126) in view of Badjatiya (US 2022/0237406).
Regarding claim 1, Ramesh discloses a computer implemented method comprising:
receiving a search query for items listed on an online marketplace, the search query including an image (see: 0020, Fig. 2, Fig. 7 (702));
analyzing the image to determine characteristics of an object in the image, using an image matching service, natural language processing and/or image processing techniques to extract visual features from the image that represent the characteristics of the object (see: 0020 (image is analyzed to determined features), 0022, 0060 (note: features are analogous to characteristics), Fig. 7 (704));
automatically generating one or more search terms, including a type of the object, based on the characteristics of the object in the image (see: 0015 (keywords associated with image data), 0025 (product categories), 0027 (keywords associated with captured image), 0032 (recommended keywords are identified from text recognition or data from the image itself; category may be provided as a potential keywork), 0062, Fig. 7 (706));
Note: recommended keywords are provided, which may include a category (i.e., type) of the product.
presenting the one or more search terms in a user interface configured to receive an edit to the one or more search terms (see: Fig. 5a (510,518, 520), 0027-0030, Fig. 7 (710, 714);
Note: Ramesh teaches that the keyword recommendations include keywords in a string that are configured to initiate a keyword search when selected by the user (0027). The alternative keyword searches are presented for selection by the user (see: Fig. 5a (510, 518-520), 0029-0030, Fig. 7 (710)), which enables refining of the initial search (see: 0030, 0015, Fig. 7 (714-718)).
transmitting, over a network, the one or more search terms to one or more servers of the online marketplace to locate items matching the one or more search terms (see: 0038 (results based on the selected keyword search string from Fig. 5(a)), 0068, Fig. 6 (606), 004-0041, Fig. 7 (716); see also: 0027 (keywords in a string that are configured to initiate a keyword search when selected by the user), 0030); and
displaying visual indications of located items matching the one or more search terms in the online marketplace (see: Fig. 5b (528a-e), 0038, Fig. 7 (718), 0069).
Though disclosing all of the above, Ramesh does not disclose where the analysis is performed by a machine learning model including a convolutional neural network trained to extract the features. Ramesh, however, discloses various techniques such as object recognition and image processing, natural language processing, and text recognition (e.g., 0022, 0032). One of ordinary skill in the art would have recognized that employing such types of techniques using trained machine learning models was well-known before the effective filing date.
For example, Badjatiya teaches an image search system that employs a trained machine learning model (see: 0058-0059, 0085, Fig. 7). Further, the trained model is used to extract visual features from the image that represent the characteristics of the object, such as by extracting features from image and text data in order to provision target images as results (see: 0022, 0038, 0040, 0042, 0067, 0070, Fig. 1, Fig. 9 (960)).
Lastly, Badjatiya teaches the machine learning model to include a convolutional neural network (see: 0088, 0091-0092, 0101).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the invention of Ramesh to have utilized the known technique for employing trained machine learning models as taught by Badjatiya in order to have enabled the invention of Ramesh to have generated target features which were more accurately tailored to user desires and which in turn provided improved image search results (see: Badjatiya: 0017).
2. The computer-implemented method of claim 1, further comprising displaying the one or more search terms proximate to the visual indications of the located items in a user interface (see: Ramesh: Fig. 5b (510, 518)).
4. The computer-implemented method of claim 1, wherein the characteristics are based on intended characteristics of an item to purchase (see: Ramesh: 0022, 0028, Fig. 2-3 (310)).
7. The computer-implemented method of claim 1, wherein the machine learning model is trained using training data that includes images uploaded to the online marketplace as part of a search query (see: Badjatiya: 0021, 0059 (source image), Fig. 6 (210), Fig. 7 (720), Fig. 9 (910)).
9. The computer-implemented method of claim 1, wherein the characteristics of the object designate a price, a brand, or a designation of luxury for the object in the image (see: Ramesh: 0019 (brand), 0024, 0069 (pricing information), Fig. 5a (Brand A)).
Note: In addition to the teachings of Ramesh, merely labeling the characteristics of the items differently from the prior art fails to impart a new and nonobvious functioning of the method step. The subjective labeling of the characteristics represents non-functional descriptive material and does not patentably distinguish the claimed invention from the prior art.
27. The computer implemented method of claim 1, wherein the convolutional neural network is trained by injecting noise into training data including images of items (see: Badjatiya: 0058-0059, 0085, 0088, 0091-0092, 0101, Fig. 7).
Regarding claims 10-11, claims 10-11 recite at least substantially similar concepts and elements as recited in claims 1-2, 7 and 4 (respectively) such that similar analysis of the claims would be readily apparent to one of ordinary skill in the art. As such, claims 10-11 are rejected under at least similar rationale.
Regarding claims 21-22 and 24, claims 21-22 and 24 recite at least substantially similar concepts and elements as recited in claims 1-2 and 4 such that similar analysis of the claims would be readily apparent to one of ordinary skill in the art. As such, claims 21-22 and 24 are rejected under at least similar rationale.
Claim(s) 3 and 23 are is/are rejected under 35 U.S.C. 103 as being unpatentable over Ramesh in view of Badjatiya as applied to claims 1 and 10 above, and further in view of Grossman (US 2018/0197223).
Regarding claim 3 and parallel claim 23, Ramesh in view of Badjatiya teaches all of the above as noted but does not teach:
receiving a user input to remove at least one of the one or more search terms; and
responsive to the user input to remove the at least one of the one or more search terms, filtering the visual indications of the located items.
To this accord, Grossman teaches image-based product identification including receiving a user input to remove at least one of the one or more search terms (see: Fig. 8 (818), 0081 (deletion of a keyword); and,
responsive to the user input to remove the at least one of the one or more search terms, filtering the visual indications of the located items (see: Fig. 8 (818 [Wingdings font/0xE0] 804-816), 0081 (server 140 may proceed to identify a new or modified set of keywords at block 804, from which new purchase options may be identified and presented to the user)).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the invention of Ramesh in view of Badjatiya to have utilized the known technique for deleting search terms as taught by Grossman in order to have enabled users to manually modify searches as desired in order to identify new purchase options to the user (see: Grossman: 0081).
Claim(s) 5, 8 and 12 are is/are rejected under 35 U.S.C. 103 as being unpatentable over Ramesh in view of Badjatiya as applied to claims 1 and 10 above, and further in view of Kale (US 2017/0193011).
Regarding claim 5, Ramesh in view of Badjatiya including applying a machine learning techniques to an image search and training the machine learning model used in image searching (see again: Badjatiya). Ramesh in view of Badjatiya, however, does not teach wherein the one or more machine learning models are trained using training data that includes images of items listed on the online marketplace, the images of the items associated with characteristics of the items, the characteristics of the items extracted from listing data. One of ordinary skill in the art, however, would have readily understood that the need for “training” machine learning models was well-established.
To this accord, Kale teaches a query system for item listings that utilizes one or more machine learning models that are trained using training data that includes images of items listed on the online marketplace (see: 0018, 0039, 0044, Fig. 5 (510, 530), Fig. 6 (610, 620), Fig. 3 (330-360)), the images of the items associated with characteristics of the items, the characteristics of the items extracted from listing data (see: 0019, 0030, 0039, 0044). In Kale, images associated with the listings in the set are fed to the trained image classifiers to identify attributes (i.e., extract characteristics) of the depicted items within the listings (i.e., listing data).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the invention of Ramesh in view of Badjatiya to have utilized the known technique for training a machine learning model using a training image set as taught by Kale in order to have enabled the machine learning techniques of Ramesh in view of Badjatiya to have better identified attributes in items and improved the quality of search results while reducing the effort in searching for items (see: Kale: 0056-0057).
8. The computer-implemented method of claim 1, wherein the one or more machine learning models are trained using user purchase history (see: Kale: 0019 (interacted with listings), 0036-0037, 0043, Fig. 4 (410)
Regarding claims 12 and 15, claims 12 and 15 recite substantially similar limitations and scope as recited in claims 5 and 8. As such, claims 12 and 15 are rejected under at least similar rationale.
Claim(s) 6 and 13 are is/are rejected under 35 U.S.C. 103 as being unpatentable over Ramesh in view of Badjatiya and Kale as applied to claims 5 and 12 above, and further in view of Lee (US 2017/0256038).
Regarding claim 6 and parallel claim 13, Ramesh in view of Badjatiya and Kale teaches all of the above as noted but does not teach adding noise to the images of the items in the training data. The use of such techniques in the realm of machine learning was notoriously well-known before the effective filing date of the invention, and would have been obvious.
For example, Lee teaches a method for image analysis (e.g., abstract, 0028) that applies the known technique of training a neural network by adding noise to images in the training data (see: 0031, 0039, 0047, 0057).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the invention of Ramesh in view of Badjatiya and Kale to have utilized the known technique for adding noise to image training data as taught by Lee in order to have prevented overfitting certain training images and thereby allowing the model of Ramesh in view of Badjatiya and Kale to have become more robust against noise (see: Lee: 0031, 0047).
Claim(s) 25 is/are rejected under 35 U.S.C. 103 as being unpatentable over Ramesh in view of Badjatiya as applied to claim 1 above, and further in view of Ouyang (US 2019/0213608).
Regarding claim 25, Ramesh in view of Badjatiya teaches all of the above as noted including search refinement and editing (see again: Ramesh) but does not teach wherein the user interface includes checkboxes indicating different types of objects for receiving the edit to the one or more search terms.
Such techniques were well-established before the effective filing date of the invention and would have been obvious.
For example, Ouyang teaches a search refinement interface that comprises checkboxes indicating different options for receiving the edit to the one or more search terms (see: Fig. 6A (610B-F), 0059). Notably, Ouyang discloses a category (i.e., object type) selector (see: Fig. 6 (610A), 0059).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the invention of Ramesh in view of Badjatiya to have utilized the well-known technique for search refinement as taught by Ouyang in order to have provided search options that enabled a user to more easily and accurately refine, limit or invoke a subset of results (see: Ouyang: 0059).
Claim(s) 26 is/are rejected under 35 U.S.C. 103 as being unpatentable over Ramesh in view of Badjatiya as applied to claim 1 above, and further in view of Bagley (US 10,997,601).
Regarding claim 26, Ramesh in view of Badjatiya teaches all of the above as noted including search refinement (see again: Ramesh) but does not teach wherein the user interface includes a drop-down menu indicating different types of objects for receiving the edit to the one or more search terms.
Such techniques were well-established before the effective filing date of the invention and would have been obvious.
For example, Bagley teaches a search refinement interface that comprises a drop-down menu indicating different options for receiving the edit to the one or more search terms (see: Fig. 5 (504), col. 10 lines 53-57, Fig. 4 (404), col. 10 lines 41-44). Notably, Bagley discloses the filter corresponds to a type of product (e.g., shirt, short, pants, etc.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have modified the invention of Ramesh in view of Badjatiya to have utilized the well-known technique for search refinement as taught by Bagley in order to have provided search filters that enabled a user to more specifically refine a search (see: Bagley: col. 10 lines 53-57).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
PTO form 892-U (previously cited in the correspondence mailed 2/12/2026) demonstrates the conventionality of deliberately introducing noise to help improve computer vision models (i.e., adding noise), including various known types of noise that may be added (e.g., Gaussian, Localvar, Poisson, Salt, Pepper, s&p, speckle) (see: Paragraphs 3-4, Images under “Implementing Noise”).
McWilliams (US 2023/0401202) discloses a search interface enabling refinements using dropdowns (e.g., 410, 415, 420) and checkboxes (e.g., 425) (see: 0042, Fig. 4, Fig. 6).
Ullman (US 2014/0067795) discloses a search interface enabling refinements using dropdowns (see: 0172, Fig. 38)
PTO form 892-V discloses a training and retraining process for deep CNN models to study field specific features representations for image-based searching (see: Section III, Fig. 1).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to WILLIAM J ALLEN whose telephone number is (571)272-1443. The examiner can normally be reached Monday-Friday, 8:00-4:00.
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, Anita Coupe can be reached at 571-270-3614. 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.
WILLIAM J. ALLEN
Primary Examiner
Art Unit 3625
/WILLIAM J ALLEN/Primary Examiner, Art Unit 3619
Appendix - Proposed Amendment
The following amendment is proposed in order expedite prosecution and direct future amendments in order to place the claims in condition for allowance. Though the Examiner has included claims to be amended or cancelled below, Applicant is advised to review the remaining claims to address any potential issues for antecedent basis or correctness. The amendment is proposed as follows:
1. (Currently Amended) A computer-implemented method comprising:
receiving, by a user device, a search query for items listed on an online marketplace, the search query including [[an]] a query image received by activating a camera feature of the user device in communication with a visual search query system, the camera feature causing a camera of the user device to capture at least one image of an item;
in response to receiving the search query through the camera feature, causing the visual search query system to analyze query image to determine characteristics of an object in the query image, using a machine learning model , wherein the machine learning model is trained using images of items listed on the online marketplace, at least a portion of the images of the items having noise injected into the images;
in response to analyzing the query image, automatically generating, by the visual search query system, one or more search terms, including a type of the object, based on the characteristics of the object in the image, wherein the generation of the search terms causes the search terms to be populated into a search interface;
causing, by the visual search query system, generation of the search interface on the user device, the search interface comprising at least one interactive element, wherein each populated search term is presented with a respective interactive element, the at least one interactive element configured to receive [[an]] one or more edits to the one or more search terms, wherein the at least one interactive element includes:
one or more checkboxes indicating different types of objects for receiving the edits to the one or more search terms; or,
a drop-down menu indicating different types of objects for receiving the edits to the one or more search terms;
transmitting, over a network, the one or more search terms to one or more servers of the online marketplace to locate items matching the one or more search terms; and
displaying, in the search interface, visual indications of the items matching the one or more search terms in the online marketplace.
5. (Currently Amended) The computer-implemented method of claim 1, wherein the machine learning model is trained
6. (Cancelled)
25. (Cancelled)
26. (Cancelled)
27. (Currently Amended) The computer implemented method of claim 1, wherein the machine learning model comprises a convolutional neural network
10. (Currently Amended) A system comprising:
a device comprising:
a memory component; and
a processing device coupled to the memory component, the processing device to perform operations comprising:
receiving a search query for items listed on an online marketplace, the search query including [[an]] a query image received by activating a camera feature of the device in communication with a visual search query system, the camera feature causing a camera of the user device to capture at least one image of an item;
in response to receiving the search query through the camera feature, causing the visual search query system to analyze query image to determine characteristics of an object in the query image, using a machine learning model , wherein the machine learning model is trained using images of items listed on the online marketplace, at least a portion of the images of the items having noise injected into the images;
in response to analyzing the query image, automatically generating, by the visual search query system, one or more search terms, including a type of the object, based on the characteristics of the object in the image, wherein the generation of the search terms causes the search terms to be populated into a search interface;
causing, by the visual search query system, generation of the search interface on the device, the search interface comprising at least one interactive element, wherein each populated search term is presented with a respective interactive element, the plurality of interactive elements configured to receive [[an]] one or more edits to the one or more search terms, wherein each populated search term is presented with a respective interactive element, the at least one interactive element configured to receive [[an]] one or more edits to the one or more search terms, wherein the at least one interactive element includes:
one or more checkboxes indicating different types of objects for receiving the edits to the one or more search terms; or,
a drop-down menu indicating different types of objects for receiving the edits to the one or more search terms;
transmitting, over a network, the one or more search terms to one or more servers of the online marketplace to locate items matching the one or more search terms; and
displaying, in the search interface, visual indications of the items matching the one or more search terms in the online marketplace.
13. (Cancelled)
21. (Currently Amended) A non-transitory computer-readable storage medium storing executable instructions, which when executed by a processing device, cause the processing device to perform operations comprising:
receiving, by a user device, a search query for items listed on an online marketplace, the search query including [[an]] a query image received by activating a camera feature of a user device in communication with a visual search query system, the camera feature causing a camera of the user device to capture at least one image of an item;
in response to receiving the search query through the camera feature, causing the visual search query system to analyze query image to determine characteristics of an object in the query image, using a machine learning model , wherein the machine learning model is trained using images of items listed on the online marketplace, at least a portion of the images of the items having noise injected into the images;
in response to analyzing the query image, automatically generating, by the visual search query system, one or more search terms, including a type of the object, based on the characteristics of the object in the image, wherein the generation of the search terms causes the search terms to be populated into a search interface;
causing, by the visual search query system, generation of the search interface on the user device, the search interface comprising at least one interactive element, wherein each populated search term is presented with a respective interactive element, the at least one interactive element configured to receive [[an]] one or more edits to the one or more search terms, wherein the at least one interactive element includes:
one or more checkboxes indicating different types of objects for receiving the edits to the one or more search terms; or,
a drop-down menu indicating different types of objects for receiving the edits to the one or more search terms;
transmitting, over a network, the one or more search terms to one or more servers of the online marketplace to locate items matching the one or more search terms; and
displaying, in the search interface, visual indications of the items matching the one or more search terms in the online marketplace.
The amendments as proposed above act to (i) confer eligibility under 35 USC 101, and (ii) render the claims unobvious over the prior art in view of their specific combination of limitations as a whole. With respect to eligibility, the specific ordered combination of elements provides an improved search refinement interface which enables converting from an image-based search to text-based refinements. This is achieved through the specific sequence of operations performed by respective components in conjunction with one another (e.g., user device, camera, visual search query system) that culminates in the generation of the search interface having a particular arrangement of interactive elements (i.e., respective of each populated search term) for populated search terms resulting from the image query. This ordered combination acts to integrate any recite exception into a practical application under at Step 2A (Prong Two) also by applying any recited exception in a meaningful way beyond generally linking the use of the judicial exception to a particular technological environment and beyond the mere instructions to implement the recited exception.
Prosecution Insights