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 .
Acknowledgments
The submission filed on 03/04/26 is acknowledged.
Status of Claims
Claims 1-20 are pending.
In the amendment filed on 03/04/26, claims 1-4, 6, 7, 11, 12, 18 and 20 were amended, and no claims were cancelled or added.
Claims 1-20 are rejected.
Response to Arguments
Regarding the rejection under 35 U.S.C. 101
Applicant’s arguments have been fully considered but are not persuasive.
The Office responds to Applicant’s arguments below. In the discussion below, headings and page numbers refer to Applicant’s Response, unless otherwise indicated.
Step 2A, Prong One (p. 9)
Applicant argues:
The features of amended independent claims 1, 11, and 12 do not recite an abstract idea according to the analysis of abstract ideas under Step 2A, Prong One. The Office asserts that original independent claims 1, 11, and 12 "cover 'certain methods of organizing human activity,' specifically, 'fundamental economic practices or principles' and/or 'commercial or legal interactions"' (id., pp. 4 and 5). However, amended independent claims 1, 11, and 12 are not directed to the abstract idea of certain methods of organizing human activity, as asserted by the Office. For example, amended independent claims 1, 11, and 12 recite specific technical operations, including dividing images into sets of image segments using image segmentation logic, where the numerical quantity of image segments is based on characteristics of the respective images, and then providing image segments to a learning model to generate a confidence score associated with authenticity. The features of amended independent claims 1, 11, and 12 are directed to specific technical image processing and machine learning operations, not to fundamental economic practices or commercial interactions. The claims do not merely recite the abstract concept of authenticating items or conducting transactions. Instead, the claims recite a specific technical approach involving image segmentation and machine learning that is distinct from any method of organizing human activity. As such, the pending claims do not recite a judicial exception and are patent eligible under Step 2A, Prong One of the USPTO's subject matter eligibility guidance (MPEP § 2106.04). (p. 9; emphasis added)
As Applicant concedes, the claims recite an "abstract concept of authenticating items or conducting transactions." The fact that the claims also recite other (non-abstract idea) elements does not erase the fact that the claims recite an abstract idea.
Further, the step of dividing images and the providing of image segments to a model are also parts of the abstract idea.
The bald recitation of machine learning ("learning") is merely a generic computer element that applies the abstract idea or generally links it to a particular technological environment or field of use.1
Step 2A, Prong Two (pp. 9-11)
Applicant argues:
The MPEP provides that an "important consideration in determining whether a claim is directed to an improvement in technology is the extent to which the claim covers a particular solution to a problem or a particular way to achieve a desired outcome, as opposed to merely claiming the idea of a solution or outcome" (id. § 2106.05(a)). Finjan, Core Wireless, as well as Enfish and McRO, confirm that software-based innovations are not abstract if they make non-abstract improvements to computer technology. For example, in McRO, the claims defined a specific way, namely use of particular rules to set morph weights and transitions through phonemes, to solve the problem of producing accurate and realistic lip synchronization and facial expressions in animated characters, and thus were not directed to an abstract idea (see e.g., id. 2106.04(a)). (p. 10)
Insofar as the instant claims cover a particular solution, the particularity is of the abstract idea.
The claims do not reflect any improvement to computer functioning or other technology. Rather, the additional elements are recited at a high level of generality, are not described, and are merely used off-the-shelf in their ordinary capacities.
Applicant further argues:
Here, the subject matter provides a technical solution for an image-based authentication system to reduce a use of computational resources, as well as improve information security, by increasing an accuracy and consistency of an authentication verification process and by preventing or reducing processing of data transactions for counterfeit items (see e.g., Application, [0016]). For example, amended independent claims 1, 11, and 12 include details related to authenticating items by dividing images into sets of image segments using image segmentation logic, where the numerical quantity of image segments is based on characteristics of the respective images, and then providing image segments to a learning model to generate a confidence score associated with authenticity. The features of amended independent claims 1, 11, and 12 recite a technical solution involving specific transformations of image data that improves the functioning of a computing system and the technical field of image-based authentication. (pp. 10-11)
Application [0016] identifies the alleged improvement as being the implementing of a learning model. This is said to reduce use of computational resources by increasing accuracy and consistency of the authenticity verification process and by preventing or reducing the processing of data transactions for counterfeit items. As indicated above, the recited "learning" is merely a generic computer element recited at a high level of generality, not described, and merely used off-the-shelf in its ordinary capacity, i.e., a paradigmatic example of 'apply it'. Again, as explained above, the dividing of images is part of the abstract idea.
Applicant further argues:
Specifically, regarding amended independent claim 1, the features of "dividing, by a computing device and using image segmentation logic, respective images of the plurality of images into sets of image segments," where "a numerical quantity of image segments in the sets of image segments is based on one or more characteristics associated with the respective images," and "generating, by the computing device and as output from a learning model, a confidence score associated with an authenticity of the item based on providing one or more image segments of the sets of image segments as input to the learning model," improve the functioning of a computing system by providing a specific technical approach to image-based authentication that involves transformation of image data. Independent claim 11 has been amended to recite features similar to those of amended independent claim 1 and improves the functioning of a computing system for at least the same reasons. Regarding amended independent claim 12, the features of dividing images into sets of image segments and generating both a binary value indicating authenticity and a confidence score, and then processing or canceling data transactions based on both values, improve the functioning of a computing system by providing a specific technical pipeline for automated transaction processing based on image-based authentication results. (p. 11)
The recitations of claims 1 and 12 cited above by Applicant do not improve the functioning of a computer. Rather, the recitations constitute merely abstract idea content together with generic computer elements ("computing device"; "learning") recited at a high level of generality, not described, and merely used off-the-shelf in their ordinary capacities, i.e., a paradigmatic example of 'apply it'.
The transformation of data is distinct from the transformation or reduction of a particular article to a different state or thing; it is the latter, not the former, that may integrate the exception into a practical application. MPEP 2106.04(d)I., see also 2106.04(c) ("The nature of the article transformed. Transformation of a physical or tangible object or substance is more likely to provide significantly more (or integrate a judicial exception into a practical application) than the transformation of an intangible concept such as a contractual obligation or mental judgment.")
Step 2B (pp. 11-14)
Applicant argues the same claimed subject matter as previously presented (see pp. 11-13, first three paragraphs of argument under Step 2B). This subject matter has already been addressed above and found to amount to an abstract idea and additional elements merely applying the abstract idea or generally linking it to a particular technological environment or field of use and as such cannot provide significantly more under Step 2B.
Applicant also argues that the claims specify the "how" of the image-based authentication process, as in DDR, such as to amount to significantly more under step 2B (p. 13). To the extent that the claims specify the "how," this is just a matter of specifying the abstract idea, together with generic computer elements that are recited at a high level of generality, are not described, and are merely used off-the-shelf in their ordinary capacities, and hence that merely apply the abstract idea or generally link it to a particular technological environment or field of use, along the same lines as explained above.
As for dependent claim 6 (see pp. 13-14), the content thereof (extracting features from image segments, and selecting image segments) is merely abstract idea together with generic computer elements ("computing device"; "learning") that are recited at a high level of generality, are not described, and are merely used off-the-shelf in their ordinary capacities. As such, any alleged improvement would be an improvement in the abstract idea.
As for dependent claim 7 (see p. 14), the content thereof ("a graduated verification approach based on confidence score thresholds") is merely abstract idea together with generic computer elements ("automatically"; “for display at the computing device”; "comprising a first control selectable"; "at the first control"; "comprising a second control selectable"; and "at the second control") that are recited at a high level of generality, are not described, and are merely used off-the-shelf in their ordinary capacities. As such, any alleged improvement would be an improvement in the abstract idea.
Regarding the rejections under 35 U.S.C. 102 and 103
Applicant’s arguments have been fully considered but are moot in view of the new combinations of references being used in the current rejections.
Claim Rejections - 35 U.S.C. § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claim 7 is rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for pre-AIA the inventor(s), at the time the application was filed, had possession of the claimed invention.
Lack of Written Description/Not in Specification
Claim 7 recites:
outputting, for display at the computing device, a first request comprising a first control selectable to verify one or more attributes of a plurality of attributes associated with the item are authentic based on the confidence score satisfying a second threshold value that is less than the first threshold value, wherein the data transaction is processed responsive to a selection at the first control; or
outputting, for display at the computing device, a second request comprising a second control selectable to verify the plurality of attributes associated with the item is authentic based on the confidence score satisfying a third threshold value that is less than the first threshold value and the second threshold value, wherein the data transaction is processed responsive to a selection at the second control.
Support in the disclosure is not found for the above-indicated recitation.
As best understood, the subject matter in the originally filed disclosure most relevant to the above limitations is in the figures, Fig. 3, elements 308-320, and in the specification, paragraphs 0069 (re a request), 0091-0093 (re based on a confidence level satisfying a threshold), 0094 (re verifying attributes), 0096 (re displaying a control), 0097 (re processing a data transaction responsive to selection at a control). While 0092 and 312 and 0093 and 314 teach that a confidence score can be less than or greater than thresholds, nothing is seen to teach or suggest that a second threshold value is less than the first threshold value, or that a third threshold value is less than the first threshold value and the second threshold value. The magnitude of the second threshold value relative to the first threshold value, and the magnitude of the third threshold value relative to the first threshold value and the second threshold value, are not taught or suggested.
Accordingly, support in the disclosure is not found for the above-indicated limitations of claim 7.
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 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Claims 1-20 are directed to a method or system, which are/is one of the statutory categories of invention. (Step 1: YES)
Claims 1, 11 and 12 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The claims recite a method and system for determining whether an item is authentic or counterfeit and (in the case of claim 12) proceeding with or cancelling a transaction, respectively (note as per the specification the item can be an item listed for sale in an online marketplace, and the transaction can be a merchant paying a payment platform or an online marketplace causing a shipper to distribute an item, see 0043-0047, 0073, 0078, 0087, 0097, 0130; see also dependent claims 9 and 10).
For claims 1 and 11 (claim 11 being deemed representative), the limitations (indicated below in bold) of:
one or more processors; and
a computer-readable storage medium storing instructions that are executable by the one or more processors to perform operations comprising:
receiving, from an image capture system, a plurality of images of an item;
dividing, by a computing device and using image segmentation logic, respective images of the plurality of images into sets of image segments, wherein a numerical quantity of image segments in the sets of image segments is based on one or more characteristics associated with the respective images;
generating, by the computing device and as output from a learning model, a confidence score associated with an authenticity of the item based on providing one or more image segments of the sets of image segments as input to the learning model; and
broadcasting, by the computing device, the confidence score for displaying the authenticity of the item via a user interface.
as drafted, constitute a process that, under the broadest reasonable interpretation, covers "certain methods of organizing human activity," specifically, "fundamental economic practices or principles" and/or "commercial or legal interactions" but for recitation of generic computer components and generally linking the use of a judicial exception to a particular technological environment or field of use. The Examiner notes that "fundamental economic practices" or "fundamental economic principles" describe concepts relating to the economy and commerce, including hedging, insurance, and mitigating risks, and "commercial interactions" or "legal interactions" include agreements in the form of contracts, legal obligations, advertising, marketing or sales activities or behaviors, and business relations, MPEP 2106.04(a)(2)II.A.,B. If a claim limitation, under its broadest reasonable interpretation, covers "fundamental economic practices or principles" and/or "commercial or legal interactions," but for recitation of generic computer components and generally linking the use of a judicial exception to a particular technological environment or field of use, then it falls within the "certain methods of organizing human activity" grouping of abstract ideas. Accordingly, claims 1 and 11 recite an abstract idea. (Step 2A - Prong 1: YES. The claims recite an abstract idea.)
For claims 12, the limitations (indicated below in bold) of:
receiving, from an image capture system, a plurality of images of an item;
dividing, by a computing device and using image segmentation logic, respective images of the plurality of images into sets of image segments, wherein a numerical quantity of image segments in the sets of image segments is based on one or more characteristics associated with the respective images;
generating, by the computing device and as output from a learning model, a binary value that indicates an authenticity of the item and a confidence score associated with the authenticity of the item based on providing one or more image segments of the sets of image segments as input to the learning model; and
processing or canceling, by the computing device, one or more data transactions associated with the item based on the binary value that indicates the authenticity of the item and the confidence score associated with the authenticity of the item.
as drafted, constitute a process that, under the broadest reasonable interpretation, covers "certain methods of organizing human activity," specifically, "fundamental economic practices or principles" and/or "commercial or legal interactions" but for recitation of generic computer components and generally linking the use of a judicial exception to a particular technological environment or field of use. The Examiner notes that "fundamental economic practices" or "fundamental economic principles" describe concepts relating to the economy and commerce, including hedging, insurance, and mitigating risks, and "commercial interactions" or "legal interactions" include agreements in the form of contracts, legal obligations, advertising, marketing or sales activities or behaviors, and business relations, MPEP 2106.04(a)(2)II.A.,B. If a claim limitation, under its broadest reasonable interpretation, covers "fundamental economic practices or principles" and/or "commercial or legal interactions," but for recitation of generic computer components and generally linking the use of a judicial exception to a particular technological environment or field of use, then it falls within the "certain methods of organizing human activity" grouping of abstract ideas. Accordingly, claim 12 recites an abstract idea. (Step 2A - Prong 1: YES. The claims recite an abstract idea.)
This judicial exception is not integrated into a practical application. Claims 1, 11 and 12 recite the additional elements of an image capture system, a computing device, learning, and a user interface (the foregoing recited by claim 1); one or more processors, a computer-readable storage medium, an image capture system, a computing device, learning, and a user interface (the foregoing recited by claim 11); and an image capture system, a computing device, and learning (the foregoing recited by claim 12), that implement the abstract idea. These additional elements are not described by the applicant and they are recited at a high level of generality (i.e., one or more generic computer elements performing generic computer functions, or generally linking the use of a judicial exception to a particular technological environment or field of use), such that they amount to no more than mere instructions to apply the exception using generic computer elements (namely, one or more processors, a computer-readable storage medium, a computing device, learning, and a user interface), or such that they amount to no more than generally linking the use of a judicial exception to a particular technological environment or field of use (namely, an image capture system and learning). Accordingly, even in combination these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. (Step 2A - prong 2: NO. The additional elements do not integrate the abstract idea into a practical application.)
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception itself. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements of an image capture system, a computing device, learning, and a user interface (the foregoing recited by claim 1); one or more processors, a computer-readable storage medium, an image capture system, a computing device, learning, and a user interface (the foregoing recited by claim 11); and an image capture system, a computing device, and learning (the foregoing recited by claim 12), to perform the noted steps amount to no more than mere instructions to apply the exception using generic computer elements or generally linking the use of a judicial exception to a particular technological environment or field of use. Mere instructions to apply an exception using generic computer elements or generally linking the use of a judicial exception to a particular technological environment or field of use cannot provide an inventive concept ("significantly more"). Accordingly, even in combination, these additional elements do not provide significantly more. As such, Claims 1, 11 and 12 are not patent eligible. (Step 2B: NO. The claims do not provide significantly more.)
Dependent claims 2-10 and 13-20 are similarly rejected because they further define/narrow the abstract idea of independent claims 1, 11 and 12 as discussed above, and/or do not integrate the abstract idea into a practical application or provide an inventive concept such as would render the claims eligible, whether each is considered individually or as an ordered combination.
As for further defining/narrowing the abstract idea:
Dependent claim 2 merely describes wherein the … model includes a feature component and a classifier component, and wherein generating the confidence score comprises: receiving, as output from the feature component of the … model, one or more feature vectors representative of the one or more image segments based on providing the sets of image segments as input to the feature component of the … model; and receiving, as output from the classifier component of the … model, the confidence score based on providing the one or more feature vectors as input to the classifier component of the … model.
Dependent claim 3 merely describes wherein the one or more feature vectors are associated with attributes of the one or more image segments.
Dependent claim 4 merely describes obtaining training data that includes an additional plurality of images of respective items for input to the … model and an additional authenticity of the respective items; and training, by minimizing a loss function using the training data, the classifier component of the … model to determine the confidence score.
Dependent claim 5 merely describes wherein the confidence score fails to satisfy a threshold value, the … method further comprising: receiving … an indication of the authenticity of the item; and retraining, by minimizing the loss function using the plurality of images of the item and the indication of the authenticity of the item, the classifier component of the … model to determine the confidence score.
Dependent claim 6 merely describes extracting, based on providing the sets of image segments as input to an additional … model, one or more features from respective image segments of the sets of image segments, and selecting the one or more image segments based on the one or more features associated with the one or more image segments being associated with the authenticity of the item.
Dependent claim 7 merely describes one of: … processing a data transaction associated with the item based on the confidence score satisfying a first threshold value; outputting … a first request … to verify one or more attributes of a plurality of attributes associated with the item are authentic based on the confidence score satisfying a second threshold value that is less than the first threshold value, wherein the data transaction is processed responsive to a selection …; or outputting … a second request … to verify the plurality of attributes associated with the item is authentic based on the confidence score satisfying a third threshold value that is less than the first threshold value and the second threshold value, wherein the data transaction is processed responsive to a selection ….
Dependent claim 8 merely describes to indicate a first value or a second value associated with a true authenticity of the item based on the confidence score failing to satisfy a threshold value.
Dependent claim 9 merely describes receiving a selection of the first value …; and processing a data transaction associated with the item based on the selection.
Dependent claim 10 merely describes receiving a selection of the second value …; and canceling processing of a data transaction associated with the item based on the selection.
Dependent claim 13 merely describes wherein processing or canceling the one or more data transactions associated with the item comprises processing the one or more data transactions based on the binary value that indicates the authenticity of the item indicating that the item is authentic and based on the confidence score satisfying one or more threshold values.
Dependent claim 14 merely describes wherein processing or canceling the one or more data transactions associated with the item comprises canceling the one or more data transactions based on the binary value that indicates the authenticity of the item indicating that the item is counterfeit and based on the confidence score satisfying one or more threshold values.
Dependent claim 15 merely describes determining the confidence score fails to satisfy at least one threshold value; … to indicate a true authenticity of the item based on the confidence score failing to satisfy the at least one threshold value; and receiving a selection.
Dependent claim 16 merely describes wherein processing or canceling the one or more data transactions associated with the item comprises processing the one or more data transactions based on the selection indicating that the true authenticity of the item is authentic.
Dependent claim 17 merely describes wherein processing or canceling the one or more data transactions associated with the item comprises canceling the one or more data transactions based on the selection indicating that the true authenticity of the item is counterfeit.
Dependent claim 18 merely describes … the … model based on the selection and the one or more image segments.
Dependent claim 19 merely describes wherein the one or more data transactions are associated with one or more of a distribution of the item or a sale of the item.
Dependent claim 20 merely describes wherein the … model includes a feature component and a classifier component, and wherein generating the binary value that indicates the authenticity of the item and the confidence score associated with the authenticity of the item comprises: receiving, as output from the feature component of the … model, one or more feature vectors representative of the one or more image segments based on providing the one or more image segments as input to the feature component of the … model; and receiving, as output from the classifier component of the … model, the binary value that indicates the authenticity of the item and the confidence score associated with the authenticity of the item based on providing the one or more feature vectors as input to the classifier component of the … model.
As for additional elements:
Claims 2-10 and 13-20 recite that the method is “computer-implemented.” This recitation is at a high level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer element. Even in combination these additional elements do not integrate the abstract idea into a practical application and do not amount to significantly more than the abstract idea itself.
Claims 2, 4-6, 18 and 20 further recite “learning.” This recitation is at a high level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer element or generally linking the use of a judicial exception to a particular technological environment or field of use. Even in combination these additional elements do not integrate the abstract idea into a practical application and do not amount to significantly more than the abstract idea itself.
Claim 5 further recites “via at least one control of the user interface.” This recitation is at a high level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer element. Even in combination these additional elements do not integrate the abstract idea into a practical application and do not amount to significantly more than the abstract idea itself.
Claim 6 further recites “by the computing device.” This recitation is at a high level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer element. Even in combination these additional elements do not integrate the abstract idea into a practical application and do not amount to significantly more than the abstract idea itself.
Claim 7 further recites "automatically," “for display at the computing device,” "comprising a first control selectable," "at the first control," "comprising a second control selectable," and "at the second control." This recitation is at a high level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer element. Even in combination these additional elements do not integrate the abstract idea into a practical application and do not amount to significantly more than the abstract idea itself.
Claim 8 further recites “causing display of a control at the user interface, wherein the control is selectable." This recitation is at a high level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer element. Even in combination these additional elements do not integrate the abstract idea into a practical application and do not amount to significantly more than the abstract idea itself.
Claims 9 and 10 further recite "… via the control." This recitation is at a high level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer element. Even in combination these additional elements do not integrate the abstract idea into a practical application and do not amount to significantly more than the abstract idea itself.
Claim 15 further recites "causing display of a control at a user interface, wherein the control is selectable …; and … via the control." This recitation is at a high level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer element. Even in combination these additional elements do not integrate the abstract idea into a practical application and do not amount to significantly more than the abstract idea itself.
Claim 18 further recites “retraining.” This recitation is at a high level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer element or generally linking the use of a judicial exception to a particular technological environment or field of use. Even in combination these additional elements do not integrate the abstract idea into a practical application and do not amount to significantly more than the abstract idea itself.
Therefore, dependent claims 2-10 and 13-20 are not patent eligible.
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 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 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.
The factual inquiries set forth in Ramasubramanian v. John Deere Co., 383 U.S. 1, 148 USPQ 459 (1966), that are applied for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-4, 6, 7, 11-14, 19 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Ramasubramanian et al. (U.S. Patent Application Publication No. 2022/0351215 A1), hereafter Ramasubramanian, in view of Asendorf et al. (U.S. Patent Application Publication No. 2020/0364513 A1), hereafter Asendorf.
Regarding Claims 1 and 11
Ramasubramanian teaches:
(claim 11) one or more processors; and (Fig. 6, 602)
(claim 11) a computer-readable storage medium storing instructions that are executable by the one or more processors to perform operations comprising: (Fig. 6, 604; claims 8 and 15)
receiving, from an image capture system, a plurality of images of an item; (0021, 0023, 0024, 0034, 0045, claims 1 and 2, online shopping marketplace / item receiver 122 / counterfeit identification model 202B / online shopping server/site receives information about item, including image data; note: images/image data that is received by a computer (such as indicated above, e.g., 122, 202B) is received from some other device, and the other device necessarily has acquired (captured) the images/image data, therefore the other device is an image capture system; regarding a plurality of images: 0023 "one or more images [of an item]… different views"; 0034, 0045 "one or more photos of the item from different views")
dividing, by a computing device (Fig. 1, 120 online shopping server; Fig. 6, 600/602, 0013) and using image segmentation logic, respective images of the plurality of images into sets of image segments (e.g., pixels), wherein … image segments in the sets of image segments … the respective images; (0025 "analyzing the image data may include a pixel level comparison of the image data of the item with one or more images of counterfeit items." -- each/any of the received images may be divided into pixels; the pixels are image segments; for each image, the set of its pixels is a set of image segments; alternative teachings include: 0026 "For example, use of the machine-learning models may include pattern-matching based on feature vectors associated with predefined portions [segments] of image data. Categories of the items may determine the predefined portion for matching. For example, the model may be trained to focus on comparing a printed logo pattern [segment] on the side view of items when the item is an accessory bag."; 0033 "Additionally or alternatively, the counterfeit identification model 202A may fine-tune the output from the decoder 208 for determining counterfeit for a task-specific dataset, which represents predetermined categories of items. … Each predetermined category may specify particular parts [segments] of the respective items to compare in finer details than other parts. For example, texture, top buttons, rivets, a waistband label, a tab label, an average market price, and the like."; 0035 "The embedding generator 210 generates embedding of the received information about the item, including the image data. In aspects, the embedding includes a multi-dimensional vector mapping of the image data. Embedding helps identify similarities of parts [segments] of the image data with parts of respective images of trained images."; regarding using image segmentation logic: 0053-0057, Fig. 6, 602, 604 (instructions for identifying counterfeit") The device that performs the method/operations (0053) includes a computer/processor 602 and instructions/program modules (0055); by virtue of performing the indicated functionality of performing a pixel level comparison (0025), the instructions carried out by the processor constitute image segmentation logic)
generating, by the computing device and as output from a learning model (Fig. 1, item matcher 124, counterfeit determiner 126, trained model 144; Fig. 2A, counterfeit identification model 202A including decoder 208; Fig. 2B, counterfeit identification model 202B including match predictor 212), a confidence score (0036-0037 "confidence value") associated with an authenticity of the item based on providing one or more image segments of the sets of image segments as input to the learning model; and (0036-0037; regarding based on providing the plurality of image segments as input to the learning model: 0024-0026 the pixels (0025), "portions"/"parts" (0026) have been provided to the learning model as input (see 0024, 0026); 0035 the "parts" have been provided to the learning model as input; claims 2, 9, 16 the pixels have been provided to the learning model as input; claims 3, 10, 17 the parts have been provided to the learning model as input; note: where an image, or image data thereof, has been provided as input to a model, one of ordinary skill understands from 0025's teaching re pixels that the pixels of the image/ image data are included in what is provided as input to the model; note the learning model is also taught by the components, collectively, of 120 in Fig. 1)
broadcasting, by the computing device, the confidence score for displaying the authenticity of the item via a user interface. (0036 under the broadest reasonable interpretation, "the match predicter 212 outputs … a confidence value" teaches broadcasting, by the computing device, the confidence score; see also 0005, 0027, 0052, Fig. 5, 516, claim 1; note "for displaying the authenticity of the item via a user interface" is intended use)
Ramasubramanian does not explicitly disclose but Asendorf teaches:
dividing, by a computing device and using image segmentation logic, … image… into set… of image segments, wherein a numerical quantity of image segments in the set… of image segments is based on one or more characteristics associated with the … image…; (0058 "At 820, one or more patches are extracted from the test sample image, similar as in 220 of FIG. 2. The number l of patches may be, for example, from one to 1,000,000. … The patches may have … a patch size, for example, from 1×1 pixel to 1000×1000 pixels. It is to be understood that the suitable number of patches and the suitable patch size may depend on properties of the sample images (e.g., size, resolution, shape, quality, etc.)." -- the patch teaches the "image segment"; where the number of patches exceeds 1, the set of patches teaches the "set of image segments"; the dependence of the number of patches on the image properties teaches "wherein a numerical quantity of image segments in the set… of image segments is based on one or more characteristics associated with the … image…"; note although Asendorf teaches dividing a single image of a test sample ("item"), Ramasubramanian teaches "a plurality of images of an item" and "dividing, by a computing device and using image segmentation logic, respective images of the plurality of images into sets of image segments, wherein … image segments in the sets of image segments … the respective images;," as indicated above; also, it would be obvious based on legal precedent to duplicate Asendorf's teachings so as to arrive at dividing multiple images into sets of image segments, i.e., to arrive at "dividing, by a computing device and using image segmentation logic, respective images of the plurality of images into sets of image segments, wherein a numerical quantity of image segments in the sets of image segments is based on one or more characteristics associated with the respective images;", MPEP 2144.04 VI.B. (Duplication of Parts))
… providing one or more image segments of the set… of image segments as input to the learning model; (0062-0063, Fig. 8, 860-880 (see more detailed account of this same teaching that is provided at 0045-0046; 0062-0063 pertain to using the model and 0045-0046 pertain to training the model): patch feature vectors are evaluated by trained algorithm, individually (870) and collectively (880), to determine if sample ("item") is authentic or counterfeit; under broadest reasonable interpretation, provision of a patch feature vector teaches provision of a patch (image segment); note, as explained for the preceding bullet point immediately above, although Asendorf teaches providing image segments of a single set of image segments, Ramasubramanian teaches "providing one or more image segments of the sets of image segments as input to the learning model," and it would be obvious based on legal precedent to duplicate Asendorf's teachings so as to arrive at providing image segments of multiple sets of image segments, i.e., to arrive at "providing one or more image segments of the sets of image segments as input to the learning model;", MPEP 2144.04 VI.B. (Duplication of Parts))
It would have been obvious to one of ordinary skill in the art not later than the effective filing date of the claimed invention to have modified Ramasubramanian's systems and methods for identifying counterfeits, by incorporating therein these teachings of Asendorf regarding dividing an image into patches (segments) based on characteristic(s) of the image, because the combination would strike the appropriate balance between accuracy and efficiency, e.g., where the image is large or complex, the image is divided into a larger number of patches to improve accuracy of image analysis (accuracy of determination as authentic or counterfeit), and where the image is small or simple, the image is divided into a smaller number of patches larger so as to minimize processing and conserve processing resources (since accuracy would not be improved to such an extent as would justify increasing the number of patches).
Note Asendorf's patches could be incorporated into Ramasubramanian as intermediate (mid-size) entities between the level of the image/image portion (large size entities) and the level of the pixel (small size entities), or alternatively Asendorf's patches could be incorporated into Ramasubramanian as small size entities in place of Ramasubramanian's pixels. According to these ways of combining, the combination would also be obvious because it is merely a matter of combining prior art elements according to known methods to yield predictable results, MPEP 2143.I.A., or alternatively because it is merely a matter of simple substitution of one known element for another to obtain predictable results, MPEP 2143.I.B.
Regarding Claim 2
Ramasubramanian in view of Asendorf teaches the limitations of base claim 1 as set forth above. Ramasubramanian further teaches:
wherein the learning model (Fig. 2A, counterfeit identification model 202A; Fig. 2B, counterfeit identification model 202B) includes a feature component (0032 feature extractor 204, Fig. 2A; 0033 encoder 206, Fig. 2A; 0035 embedding generator 210, Fig. 2B) and a classifier component (Fig. 1, item matcher 124, counterfeit determiner 126, and trained model 144, collectively; Fig. 2A, decoder 208; Fig. 2B, match predictor 212 and match provider 214, collectively), and wherein generating the confidence score (0036-0037 "confidence value") comprises: (note the learning model is also taught by the components, collectively, of 120 in Fig. 1)
receiving, as output from the feature component of the learning model, one or more feature vectors representative of the one or more image segments based on providing the sets of image segments as input to the feature component of the learning model; and (0025-0026 0032-0033, 0035, 0046, Fig. 5, 506 e.g., extract features, generate feature vectors; note, under broadest reasonable interpretation, inasmuch as an image is made up of its pixels, if a feature vector is representative of the image then by definition the feature vector is representative of the image's pixels)
receiving, as output (0036, 0005, 0027, 0052, Fig. 5, 516, claim 1) from the classifier component of the learning model, the confidence score (0036-0037 "confidence value") based on providing the one or more feature vectors as input to the classifier component of the learning model. (0005, 0027, 0052, Fig. 5, 516, claim 1, 0036-0037; regarding based on providing the one or more feature vectors as input to the classifier component of the learning model: 0024-0026, 0032-0037, 0046-0049, Fig. 5, claims 2 and 3, the features (0025-0026, 0032-0033, 0035, 0046, Fig. 5, 506, claims 2 and 3) have been provided as input to the classifier component of the learning model)
Alternatively, Asendorf further teaches:
receiving, as output from the feature component of the learning model, one or more feature vectors representative of the one or more image segments based on providing the set… of image segments as input to the feature component of the learning model; and (0029 "At 120, the digital images are processed, via the processor, to extract computer-vision features."; 0060-0061, Fig. 8, 840, 850 generate patch feature vectors (feature vectors representative of the one or more image segments) from patches (image segments) (see more detailed account of this same teaching that is provided at 0040-0043; 0060-0061 pertain to using the model and 0040-0043 pertain to training the model); regarding the feature component of the learning model: 0075, 0077, Fig. 9, processor 912 extracts the features from the images and accordingly teaches "the feature component of the learning model"; note, as explained at the rejection of claims 1 and 11 above, although Asendorf teaches a single set of image segments for a single image of an item, Ramasubramanian teaches multiple sets of image segments for multiple images of an item, and it would be obvious based on legal precedent to duplicate Asendorf's teachings so as to arrive at multiple sets of image segments for multiple images of an item, i.e., to arrive at "receiving, as output from the feature component of the learning model, one or more feature vectors representative of the one or more image segments based on providing the sets of image segments as input to the feature component of the learning model;," MPEP 2144.04 VI.B. (Duplication of Parts))
Regarding Claim 3
Ramasubramanian in view of Asendorf teaches the limitations of base claim 1 and intervening claim 2 as set forth above. Ramasubramanian further teaches:
wherein the one or more feature vectors are associated with attributes of the one or more image segments. (0025-0026 0032-0033, 0035, 0046 attribute is taught by any of category, short description, pattern (0026), edge, color, metadata (0032 teaches various examples thereof), part specified by category (0033); note, under broadest reasonable interpretation, inasmuch as an image is made up of its pixels, if an attribute pertains to (or: is of) an image then by definition the attribute pertains to (or: is of) the image's pixels)
Alternatively, Asendorf further teaches:
wherein the one or more feature vectors are associated with attributes of the one or more image segments. (0029 "At 120, the digital images are processed, via the processor, to extract computer-vision features. The computer-vision features can represent characteristic features of the material samples. For example, a digital image of a material sample may include pixels having various intensities/colors which may be related to a structure, a substructure, or a texture of the material sample. In some embodiments, processing the image of a respirator can extract computer-vision features representing a texture of the respirator material (e.g., a nonwoven material). The characteristic features of a material sample may include, for example, an optical feature (e.g., intensity, color, etc.), a surface pattern feature, an acoustical feature (e.g., frequency absorption), an elastic feature (e.g., modulus), a structural feature (e.g., shape), an electronic feature (e.g., resistance), a magnetic feature (e.g., field strength), an electrets related feature (e.g., dielectric), a mechanical feature (e.g., yield strength), etc." -- note the preceding underlined items represent attributes not merely of the item but also of the image (attributes of the one or more image segments); see also 0060-0061, Fig. 8, 840, 850, and the more detailed account of this same teaching that is provided at 0040-0043, where LBP (0040) is explained at 0038 and the filtering is explained at 0036-0037, both of which explanations teach that the one or more feature vectors are associated with attributes of the one or more image segments (0060-0061 pertain to using the model and 0036-0043 pertain to training the model))
Regarding Claim 4
Ramasubramanian in view of Asendorf teaches the limitations of base claim 1 and intervening claim 2 as set forth above. Ramasubramanian further teaches:
obtaining training data that includes an additional plurality of images of respective items for input to the learning model and an additional authenticity of the respective items; and (0028-0029; regarding an additional authenticity of the respective items: "image data of items that have been confirmed as counterfeit in the past" (0029); regarding "additional": (note the term "additional" is relative to the "plurality of images" and "authenticity" recited in claim 1) The training data (the images that are inputted, and the authenticity that is outputted) used for training the model in 0028-0029 is different from the actual image data that is inputted to the trained model in order to determine the authenticity/counterfeit status thereof (e.g., this actual image input data is taught, e.g., at 0024-0026, 0035, claims 2, 9, 16, claims 3, 10, 17, as per claims 1 and 11, "generating" step, above) and different from the authenticity/ counterfeit status of the item that is outputted from the trained model based on the input to the trained model (e.g., this output data regarding the item is taught, e.g., at 0036-0037, as per claims 1 and 11, "generating" step, above, see also 0005, 0027, 0049, Fig. 5, 510); therefore, the training data (images and authenticity) is additional)
training, … using the training data, the classifier component of the learning model to determine the confidence score. (0028-0029; regarding the confidence score: 0036-0037)
Asendorf further teaches:
by minimizing a loss function …. (0046 "A suitable classification algorithm can be selected to provide an accurate prediction/identification for test samples, to minimize training error, or minimize error on a validation dataset.")
It would have been obvious to one of ordinary skill in the art not later than the effective filing date of the claimed invention to have modified the combination of Ramasubramanian's systems and methods for identifying counterfeits, as modified by Asendorf's teachings regarding dividing an image into patches (segments) based on characteristic(s) of the image, by incorporating therein these further teachings of Asendorf's teachings regarding training to minimize an error (loss function), because minimizing a loss function is a standard, fundamental aspect of training machine learning, which serves to improve accuracy of the results, and as such the combination is merely a matter of combining prior art elements according to known methods to yield predictable results; use of known technique to improve similar devices (methods, or products) in the same way; and applying a known technique to a known device (method, or product) ready for improvement to yield predictable results. MPEP 2143.I.A.,C.,D.
Regarding Claim 6
Ramasubramanian in view of Asendorf teaches the limitations of base claim 1 as set forth above. Ramasubramanian further teaches:
extracting, by the computing device and based on providing the sets of image segments as input to an additional learning model, one or more features from respective image segments of the sets of image segments; and (0025-0026 0032-0033, 0035, 0046, Fig. 5, 506 e.g., extract features, generate feature vectors; note, under broadest reasonable interpretation of the term "respective image segments," inasmuch as the features are extracted from particular parts of an image (e.g., product label/serial number/side of item (0025), predefined portions/logo pattern/distinct parts (0026), edge information (0032), particular parts (0033)), the features are extracted from the particular sets of pixels respectively constituting those particular parts, thus teaching extracting … one or more features from respective image segments of the sets of image segments; regarding an additional learning model: (note the term "additional" is relative to the "learning model" recited in claim 1, "generating" step) The learning model used to extract the features is described as a "machine-learning model" (0026), whereas the learning model of clam 1, "generating" step, is described as a "neural network," of various types (0036), thus teaching or suggesting that the former is different from the latter, i.e., teaching "an additional learning model"; alternatively, it would be obvious based on legal precedent to duplicate Ramasubramanian's machine learning model so as to arrive at "an additional learning model," MPEP 2144.04 VI.B. (Duplication of Parts))
selecting the one or more image segments based on the one or more features associated with the one or more image segments being associated with the authenticity of the item. (0025 "Features of the image data for comparison may depend on a category of an item. For example, when a category of the item is electronic devices, features of the image data for the item may include a photo of a product label with a product serial number. In another example, when a category of the item is an accessory bag, the feature may include a close-up photo or printed patterns on the side of the accessory bag." ; 0026 "For example, use of the machine-learning models may include pattern-matching based on feature vectors associated with predefined portions of image data. Categories of the items may determine the predefined portion for matching. For example, the model may be trained to focus on comparing a printed logo pattern on the side view of items when the item is an accessory bag."; 0033 "Additionally or alternatively, the counterfeit identification model 202A may fine-tune the output from the decoder 208 for determining counterfeit for a task-specific dataset, which represents predetermined categories of items. For example, the predetermined categories of items may include handbag accessories, shoes, and pairs of vintage jeans. Each predetermined category may specify particular parts of the respective items to compare in finer details than other parts. For example, texture, top buttons, rivets, a waistband label, a tab label, an average market price, and the like."; note, under broadest reasonable interpretation, inasmuch as an image is made up of its pixels, selecting a particular part of an image teaches selecting the pixels (image segments) of that particular part of the image)
Alternatively, Asendorf further teaches:
extracting, by the computing device and based on providing the set… of image segments as input to an additional learning model, one or more features from respective image segments of the set… of image segments; and (0029; 0060-0061, Fig. 8, 840, 850 generate patch feature vectors (feature vectors representative of the one or more image segments) from patches (image segments) (see more detailed account of this same teaching that is provided at 0040-0043; 0060-0061 pertain to using the model and 0040-0043 pertain to training the model); regarding an additional learning model: 0075, 0077, Fig. 9, processor 912 extracts the features from the images and accordingly in this functionality teaches "an additional learning model"; ; alternatively, it would be obvious based on legal precedent to duplicate Asendorf's machine learning model so as to arrive at "an additional learning model," MPEP 2144.04 VI.B. (Duplication of Parts); note, as explained at the rejection of claims 1 and 11 above, although Asendorf teaches a single set of image segments for a single image of an item, Ramasubramanian teaches multiple sets of image segments for multiple images of an item, and it would be obvious based on legal precedent to duplicate Asendorf's teachings so as to arrive at multiple sets of image segments for multiple images of an item, i.e., to arrive at "extracting, by the computing device and based on providing the sets of image segments as input to an additional learning model, one or more features from respective image segments of the sets of image segments," MPEP 2144.04 VI.B. (Duplication of Parts))
selecting the one or more image segments based on the one or more features associated with the one or more image segments being associated with the authenticity of the item. (0062-0063 the use of the patches to make the determination of authenticity or counterfeit status indicates that the patches have been selected for evaluation based on the fact that their features are associated with authenticity; 0055-0056 features having higher importance weights (relative to the ascertaining of authenticity/counterfeit status) are selected for use in the determination of authenticity or counterfeit status, hence the patches having those features are selected based on the fact that the features are associated with authenticity (regardless of whether all or only some of the patches actually have those features and hence are selected))
Regarding Claim 7
Ramasubramanian in view of Asendorf teaches the limitations of base claim 1 as set forth above. Ramasubramanian further teaches:
one of: automatically processing a data transaction associated with the item based on the confidence score satisfying a first threshold value; outputting, for display at the computing device, a first request comprising a first control selectable to verify one or more attributes of a plurality of attributes associated with the item are authentic based on the confidence score satisfying a second threshold value that is less than the first threshold value, wherein the data transaction is processed responsive to a selection at the first control; or outputting, for display at the computing device, a second request comprising a second control selectable to verify the plurality of attributes associated with the item is authentic based on the confidence score satisfying a third threshold value that is less than the first threshold value and the second threshold value, wherein the data transaction is processed responsive to a selection at the second control. (The following teachings teach "automatically processing a data transaction associated with the item based on the confidence score satisfying a first threshold value": 0051, Fig. 5, 514, system updates a transaction status of the seller, e.g., system suspends seller from engaging in further transactions when it determines that item offered by seller is counterfeit. See Fig. 4, 404, 0042-0043 as to the seller status. Under broadest reasonable interpretation, the updating of status teaches the processing a data transaction as claimed; regarding based on the confidence score satisfying a threshold value: 0036-0037)
Regarding Claim 12
Ramasubramanian teaches:
receiving, from an image capture system, a plurality of images of an item;2 (0021, 0023, 0024, 0034, 0045, claims 1 and 2, online shopping marketplace / item receiver 122 / counterfeit identification model 202B / online shopping server/site receives information about item, including image data; note: images/image data that is received by a computer (such as indicated above, e.g., 122, 202B) is received from some other device, and the other device necessarily has acquired (captured) the images/image data, therefore the other device is an image capture system; regarding a plurality of images: 0023 "one or more images [of an item]… different views"; 0034, 0045 "one or more photos of the item from different views")
dividing, by a computing device (Fig. 1, 120 online shopping server; Fig. 6, 600/602, 0013) and using image segmentation logic, respective images of the plurality of images into sets of image segments (e.g., pixels), wherein … image segments in the sets of image segments … the respective images;3 (0025 "analyzing the image data may include a pixel level comparison of the image data of the item with one or more images of counterfeit items." -- each/any of the received images may be divided into pixels; the pixels are image segments; for each image, the set of its pixels is a set of image segments; alternative teachings include: 0026 "For example, use of the machine-learning models may include pattern-matching based on feature vectors associated with predefined portions [segments] of image data. Categories of the items may determine the predefined portion for matching. For example, the model may be trained to focus on comparing a printed logo pattern [segment] on the side view of items when the item is an accessory bag."; 0033 "Additionally or alternatively, the counterfeit identification model 202A may fine-tune the output from the decoder 208 for determining counterfeit for a task-specific dataset, which represents predetermined categories of items. … Each predetermined category may specify particular parts [segments] of the respective items to compare in finer details than other parts. For example, texture, top buttons, rivets, a waistband label, a tab label, an average market price, and the like."; 0035 "The embedding generator 210 generates embedding of the received information about the item, including the image data. In aspects, the embedding includes a multi-dimensional vector mapping of the image data. Embedding helps identify similarities of parts [segments] of the image data with parts of respective images of trained images."; regarding using image segmentation logic: 0053-0057, Fig. 6, 602, 604 (instructions for identifying counterfeit") The device that performs the method/operations (0053) includes a computer/processor 602 and instructions/program modules (0055); by virtue of performing the indicated functionality of performing a pixel level comparison (0025), the instructions carried out by the processor constitute image segmentation logic)
generating, by the computing device and as output from a learning model (Fig. 1, item matcher 124, counterfeit determiner 126, trained model 144; Fig. 2A, decoder 208; Fig. 2B, counterfeit identification model 202B including match predictor 212), a binary value that indicates an authenticity of the item and a confidence score (0036-0037 "confidence value") associated with the authenticity of the item based on providing the plurality of image segments as input to the learning model; and (0036-0037; regarding a binary value that indicates an authenticity of the item: the language "may determine that the item X is counterfeit" and "may determine that the item Y is not a counterfeit item" (0037) and "may indicate a counterfeit item X" (0036) indicates that the system is making a yes/no (binary) decision as to whether the item is authentic (or equivalently, that the system is making a yes/no (binary) decision as to whether the item is counterfeit, or equivalently, that the system is making a (binary) decision as to whether the item is authentic or counterfeit); therefore, this teaches a binary value that indicates an authenticity of the item; note the "confidence value" (confidence score) is generated and outputted as an item distinct from the binary value that indicates authenticity of the item, e.g., the "confidence value" (confidence score) may be generated and outputted as a percentage, e.g., "an output may indicate a counterfeit item X [the binary value that indicates authenticity of the item] with a confidence value [confidence score] or [sic, of] 65%" (0036); regarding based on providing the plurality of image segments as input to the learning model: 0024-0026 the pixels (0025), "portions"/"parts" (0026) have been provided to the learning model as input (see 0024, 0026); 0035 the "parts" have been provided to the learning model as input; claims 2, 9, 16 the pixels have been provided to the learning model as input; claims 3, 10, 17 the parts have been provided to the learning model as input; note: where an image, or image data thereof, has been provided as input to a model, one of ordinary skill understands from 0025's teaching re pixels that the pixels of the image/image data are included in what is provided as input to the model; note the learning model is also taught by the components, collectively, of 120 in Fig. 1)
processing or canceling, by the computing device, one or more data transactions associated with the item based on the binary value that indicates the authenticity of the item and the confidence score associated with the authenticity of the item. (0005, 0023, 0027 "disables transactions associated with the item in the online shopping site. For example, the transaction disabler 128 may disable sales of the item when the item is already on listings on the online shopping site. Additionally or alternatively, the transaction disabler 128 may disable listing of the items that the seller has previously posted on the online shopping site for sales transaction."; 0051, claims 1 and 3; note: the claim as drafted ("processing or canceling") does not require that the prior art teach both processing and canceling)
Ramasubramanian does not explicitly disclose but Asendorf teaches:
dividing, by a computing device and using image segmentation logic, … image… into set… of image segments, wherein a numerical quantity of image segments in the set… of image segments in the set… of image segments is based on one or more characteristics associated with the … image…;4 (0058 "At 820, one or more patches are extracted from the test sample image, similar as in 220 of FIG. 2. The number l of patches may be, for example, from one to 1,000,000. … The patches may have … a patch size, for example, from 1×1 pixel to 1000×1000 pixels. It is to be understood that the suitable number of patches and the suitable patch size may depend on properties of the sample images (e.g., size, resolution, shape, quality, etc.)." -- the patch teaches the "image segment"; where the number of patches exceeds 1, the set of patches teaches the "set of image segments"; the dependence of the number of patches on the image properties teaches "wherein a numerical quantity of image segments in the set… of image segments is based on one or more characteristics associated with the … image…"; note although Asendorf teaches dividing a single image of a test sample ("item"), Ramasubramanian teaches "a plurality of images of an item" and "dividing, by a computing device and using image segmentation logic, respective images of the plurality of images into sets of image segments, wherein … image segments in the sets of image segments … the respective images;," as indicated above; also, it would be obvious based on legal precedent to duplicate Asendorf's teachings so as to arrive at dividing multiple images into sets of image segments, i.e., to arrive at "dividing, by a computing device and using image segmentation logic, respective images of the plurality of images into sets of image segments, wherein a numerical quantity of image segments in the sets of image segments is based on one or more characteristics associated with the respective images;", MPEP 2144.04 VI.B. (Duplication of Parts))
… providing one or more image segments of the set… of image segments as input to the learning model;5 (0062-0063, Fig. 8, 860-880 (see more detailed account of this same teaching that is provided at 0045-0046; 0062-0063 pertain to using the model and 0045-0046 pertain to training the model): patch feature vectors are evaluated by trained algorithm, individually (870) and collectively (880), to determine if sample ("item") is authentic or counterfeit; under broadest reasonable interpretation, provision of a patch feature vector teaches provision of a patch (image segment); note, as explained for the preceding bullet point immediately above, although Asendorf teaches providing image segments of a single set of image segments, Ramasubramanian teaches "providing one or more image segments of the sets of image segments as input to the learning model," and it would be obvious based on legal precedent to duplicate Asendorf's teachings so as to arrive at providing image segments of multiple sets of image segments, i.e., to arrive at "providing one or more image segments of the sets of image segments as input to the learning model;", MPEP 2144.04 VI.B. (Duplication of Parts))
It would have been obvious to one of ordinary skill in the art not later than the effective filing date of the claimed invention to have modified Ramasubramanian's systems and methods for identifying counterfeits, by incorporating therein these teachings of Asendorf regarding dividing an image into patches (segments) based on characteristic(s) of the image, because the combination would strike the appropriate balance between accuracy and efficiency, e.g., where the image is large or complex, the image is divided into a larger number of patches to improve accuracy of image analysis (accuracy of determination as authentic or counterfeit), and where the image is small or simple, the image is divided into a smaller number of patches larger so as to minimize processing and conserve processing resources (since accuracy would not be improved to such an extent as would justify increasing the number of patches).
Note Asendorf's patches could be incorporated into Ramasubramanian as intermediate (mid-size) entities between the level of the image/image portion (large size entities) and the level of the pixel (small size entities), or alternatively Asendorf's patches could be incorporated into Ramasubramanian as small size entities in place of Ramasubramanian's pixels. According to these ways of combining, the combination would also be obvious because it is merely a matter of combining prior art elements according to known methods to yield predictable results, MPEP 2143.I.A., or alternatively because it is merely a matter of simple substitution of one known element for another to obtain predictable results, MPEP 2143.I.B.
Regarding Claim 13
Ramasubramanian in view of Asendorf teaches the limitations of base claim 12 as set forth above. Ramasubramanian further teaches:
wherein processing or canceling the one or more data transactions associated with the item comprises processing the one or more data transactions based on the binary value that indicates the authenticity of the item indicating that the item is authentic and based on the confidence score satisfying one or more threshold values. (0051, Fig. 5, 514, system updates a transaction status of the seller. See Fig. 4, 404, 0042-0043 as to the seller status ("valid" or "suspended"). Under broadest reasonable interpretation, the updating of status teaches the processing a data transaction as claimed; regarding based on the binary value that indicates the authenticity of the item indicating that the item is authentic and based on the confidence score satisfying one or more threshold values: 0036-0037; and note in particular regarding "indicating that the item is authentic": while the labels in the blocks of Fig. 5 illustrate the case of a determination that an item is counterfeit, the text description of Fig. 5 indicates that the process can also result in a determination that an item is authentic (= not counterfeit), e.g., 0047 "the trained counterfeit identification model takes information about the item as input and generates a probability distribution that indicates likelihoods of the item being counterfeit based on matching with image data of true counterfeit items." -- i.e., the resulting probability distribution is not limited to a finding of counterfeit; e.g., 0049 "the determine operation 510 compares a probability distribution as output from the matching operation. When the probability distribution data indicates a likelihood that is higher than a predetermined threshold, the determine operation 510 determines the item as not-for-sale (e.g., counterfeit)." -- i.e., it is possible for the probability distribution data to indicate a likelihood that is lower than a predetermined threshold, and the determine operation to determine the item as not counterfeit; and note in particular regarding "based on the confidence score satisfying one or more threshold values": since in a binary result determination process either of the two possible results (e.g., true or false) could be "desired" in some sense or other, it follows that a threshold may be deemed "satisfied" by reaching a threshold or by going beyond the threshold in either direction, that is, e.g., if "< 90%" indicates a binary value (result) of 0 and "= or > 90%" indicates a binary value (result) of 1, then a result of 50% satisfies the 90% threshold for the binary value (result) of 0, and a result of 95% satisfies the 90% threshold for the binary value (result) of 1; accordingly, under broadest reasonable interpretation, "satisfying" a threshold is taught by reaching a threshold or going beyond the threshold in either direction; on this interpretation, e.g., 0037 "determine that the confidence value is less than a predetermined threshold" (which yields an output of "not counterfeit") teaches satisfying a threshold.)
Regarding Claim 14
Ramasubramanian in view of Asendorf teaches the limitations of base claim 12 as set forth above. Ramasubramanian further teaches:
wherein processing or canceling the one or more data transactions associated with the item comprises canceling the one or more data transactions based on the binary value that indicates the authenticity of the item indicating that the item is counterfeit and based on the confidence score satisfying one or more threshold values. (0005, 0023, 0027 "disables transactions associated with the item in the online shopping site. For example, the transaction disabler 128 may disable sales of the item when the item is already on listings on the online shopping site. Additionally or alternatively, the transaction disabler 128 may disable listing of the items that the seller has previously posted on the online shopping site for sales transaction."; 0051, claims 1 and 3; regarding based on the binary value that indicates the authenticity of the item indicating that the item is counterfeit and based on the confidence score satisfying one or more threshold values: 0036-0037)
Regarding Claim 19
Ramasubramanian in view of Asendorf teaches the limitations of base claim 12 as set forth above. Ramasubramanian further teaches:
wherein the one or more data transactions are associated with one or more of a distribution of the item or a sale of the item.(0005, 0023, 0027 "disables transactions associated with the item in the online shopping site. For example, the transaction disabler 128 may disable sales of the item when the item is already on listings on the online shopping site. Additionally or alternatively, the transaction disabler 128 may disable listing of the items that the seller has previously posted on the online shopping site for sales transaction."; 0051, claims 1 and 3)
Regarding Claim 20
Ramasubramanian in view of Asendorf teaches the limitations of base claim 12 as set forth above. Ramasubramanian further teaches:
wherein the learning model (Fig. 2A, counterfeit identification model 202A; Fig. 2B, counterfeit identification model 202B) includes a feature component (0032 feature extractor 204, Fig. 2A; 0033 encoder 206, Fig. 2A; 0035 embedding generator 210, Fig. 2B) and a classifier component (Fig. 1, item matcher 124, counterfeit determiner 126, and trained model 144, collectively; Fig. 2A, decoder 208; Fig. 2B, match predictor 212 and match provider 214, collectively), and wherein generating the binary value that indicates the authenticity of the item and the confidence score associated with the authenticity of the item (0036-0037) comprises: (note the learning model is also taught by the components, collectively, of 120 in Fig. 1)
receiving, as output from the feature component of the learning model, one or more feature vectors representative of the one or more image segments based on providing the one or more image segments as input to the feature component of the learning model; and (0025-0026 0032-0033, 0035, 0046, Fig. 5, 506 extract features, generate feature vectors; note, under broadest reasonable interpretation, inasmuch as an image is made up of its pixels, if a feature vector is representative of the image then by definition the feature vector is representative of the image's pixels)
receiving, as output (0036, 0005, 0027, 0052, Fig. 5, 516, claim 1) from the classifier component of the learning model, the binary value that indicates the authenticity of the item and the confidence score (0036-0037 "confidence value") associated with the authenticity of the item based on providing the one or more feature vectors as input to the classifier component of the learning model. (0005, 0027, 0052, Fig. 5, 516, claim 1, 0036-0037; regarding based on providing the one or more feature vectors as input to the classifier component of the learning model: 0024-0026, 0032-0037, 0046-0049, Fig. 5, claims 2 and 3, the features (0025-0026, 0032-0033, 0035, 0046, Fig. 5, 506, claims 2 and 3) have been provided as input to the classifier component of the learning model)
Alternatively, Asendorf further teaches:
receiving, as output from the feature component of the learning model, one or more feature vectors representative of the one or more image segments based on providing the one or more image segments as input to the feature component of the learning model; and (0029 "At 120, the digital images are processed, via the processor, to extract computer-vision features."; 0060-0061, Fig. 8, 840, 850 generate patch feature vectors (feature vectors representative of the one or more image segments) from patches (image segments) (see more detailed account of this same teaching that is provided at 0040-0043; 0060-0061 pertain to using the model and 0040-0043 pertain to training the model); regarding the feature component of the learning model: 0075, 0077, Fig. 9, processor 912 extracts the features from the images and accordingly teaches "the feature component of the learning model")
Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Ramasubramanian et al. (U.S. Patent Application Publication No. 2022/0351215 A1), hereafter Ramasubramanian, in view of Asendorf et al. (U.S. Patent Application Publication No. 2020/0364513 A1), hereafter Asendorf, further in view of Cella et al. (WO2024186954A2), hereafter Cella, and further in view of Frisbee et al. (U.S. Patent Application Publication No. 2022/0036371 A1), hereafter Frisbee.
Regarding Claim 5
Ramasubramanian in view of Asendorf teaches the limitations of base claim 1 and intervening claims 2 and 4 as set forth above. Ramasubramanian further teaches:
wherein the confidence score fails to satisfy a threshold value, the computer-implemented method further comprising: (0036-0037)
receiving, … an indication of the authenticity of the item; and (0028-0029)
… training, … using the plurality of images of the item and the indication of the authenticity of the item, the classifier component of the learning model to determine the confidence score. (0028-0029)
Ramasubramanian in view of Asendorf does not explicitly disclose but Cella teaches:
retraining (0709; 0161), by minimizing the loss function (0161, see 0160-0165 for context) using the plurality of images of the item (0087, 0125, 0202, 0312-0313) and the indication of the authenticity of the item (0312-0313, 0357, 0406), the classifier component of the learning model (Abstract, 0128 "system may include a data classification module configured to classify data into classified data"; 0128 "machine learning module 312 may define one or more machine learning models … machine learning models may perform classification"; 0709 "graph neural network … classifications") to determine the confidence score (0128 "confidence score"). (0161, see 0160-0165 for context; regarding the indication of the authenticity of the item: as per 0312-0313, 0357, 0406, the labeled training data taught in 0161, 0709 may include indications of authenticity of the item; regarding the plurality of images of the item: 0087, 0125, 0202, 0312-0313 Cella's teachings are applicable to image data, e.g., 0313 "Risk identification of visual training sets (e.g., images,…)")6
It would have been obvious to one of ordinary skill in the art not later than the effective filing date of the claimed invention to have modified the combination of Ramasubramanian's systems and methods for identifying counterfeits, as modified by Asendorf's teachings regarding dividing an image into patches (segments) based on characteristic(s) of the image and regarding training to minimize an error (loss function), by incorporating therein these teachings of Cella regarding retraining to minimize a loss function, because this is a standard, fundamental aspect of machine learning/machine learning training, which serves to improve accuracy of the model and hence of the model output/results, see Cella, 0160-0165, and as such the combination is merely a matter of combining prior art elements according to known methods to yield predictable results; use of known technique to improve similar devices (methods, or products) in the same way; and applying a known technique to a known device (method, or product) ready for improvement to yield predictable results. MPEP 2143.I.A.,C.,D.
Ramasubramanian in view of Asendorf and Cella does not explicitly disclose but Frisbee teaches:
receiving, via at least one control of the user interface, an indication of the authenticity of the item; and (regarding via at least one control of the user interface: 0072; regarding receiving an indication of the authenticity of the item: per 0100/Fig. 5, the 'designating as authentic' taught by 0072 occurs in a context where the human grader can designate an item as authentic or counterfeit (an indication of the authenticity of the item) ("confirmed or rejected by … human graders"))7
It would have been obvious to one of ordinary skill in the art not later than the effective filing date of the claimed invention to have modified the combination of Ramasubramanian's systems and methods for identifying counterfeits, as modified by Asendorf's teachings regarding dividing an image into patches (segments) based on characteristic(s) of the image and regarding training to minimize an error (loss function), and as further modified by Cella's teachings regarding retraining to minimize a loss function, by incorporating therein these teachings of Frisbee regarding receiving input (human grading/evaluation re authenticity) via a UI (a UI being a standard mechanism for receiving input), because it would permit the receipt of human feedback regarding instances that are difficult for the ML model to classify, which feedback can be used to retrain the ML model, which retraining is a standard, fundamental aspect of training machine learning and serves to improve accuracy of the model and hence of the model output/results, see Frisbee, 0070-0071, 0098, see Cella, 0160-0165, and as such the combination is merely a matter of combining prior art elements according to known methods to yield predictable results; use of known technique to improve similar devices (methods, or products) in the same way; and applying a known technique to a known device (method, or product) ready for improvement to yield predictable results. MPEP 2143.I.A.,C.,D.
Claims 8 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Ramasubramanian et al. (U.S. Patent Application Publication No. 2022/0351215 A1), hereafter Ramasubramanian, in view of Asendorf et al. (U.S. Patent Application Publication No. 2020/0364513 A1), hereafter Asendorf, and further in view of Frisbee et al. (U.S. Patent Application Publication No. 2022/0036371 A1), hereafter Frisbee.
Regarding Claim 8
Ramasubramanian in view of Asendorf teaches the limitations of base claim 1 as set forth above. Ramasubramanian in view of Asendorf does not explicitly disclose but Frisbee teaches:
causing display of a control at the user interface, wherein the control is selectable to indicate a first value or a second value associated with a true authenticity of the item (regarding causing display of a control at the user interface, wherein the control is selectable to indicate a first value … associated with a true authenticity of the item: 0072 UI is presented for providing inputs, including designating card (0001, 0003, 0005 e.g., game card, whose authenticity is being evaluated) as authentic; regarding to indicate a first value or a second value associated with a true authenticity of the item: per 0100/Fig. 5, the 'designating as authentic' taught by 0072 occurs in a context where the human grader can designate an item as authentic (a first value) or counterfeit (a second value) ("confirmed (a first value) or rejected (a second value) by … human graders"))
based on the confidence score failing to satisfy a threshold value. (0119, Fig. 8, decision block 412, "No" branch; "Back at 412, the process queries whether the confidence factor is above a threshold at 412 and if the answer is no, …. If the analyzed card was not previously graded, it can be graded as discussed elsewhere [i.e., in the descriptions of Fig. 5 and 6] herein." -- that is to say, the grading of Fig. 5 or 6, which teaches to indicate a first value or a second value associated with a true authenticity of the item, is performed following (based on) the process of Fig. 8, in which it is determined that the confidence score fails to satisfy a threshold value)
It would have been obvious to one of ordinary skill in the art not later than the effective filing date of the claimed invention to have modified the combination of Ramasubramanian's systems and methods for identifying counterfeits, as modified by Asendorf's teachings regarding dividing an image into patches (segments) based on characteristic(s) of the image, by incorporating therein these teachings of Frisbee regarding performing a human review of an ML model's output of a grade/evaluation of an object for authenticity, following (based on) a confidence value failing to satisfy a threshold value, because it would increase accuracy of the output, see Frisbee, 0073, and can be used to improve the ML model by feeding back the human-determined result to retrain the model, see Frisbee, 0070.
Regarding Claim 15
Ramasubramanian in view of Asendorf teaches the limitations of base claim 12 as set forth above. Ramasubramanian in view of Asendorf does not explicitly disclose but Frisbee teaches:
determining the confidence score fails to satisfy at least one threshold value; (0119, Fig. 8, decision block 412, "No" branch; "Back at 412, the process queries whether the confidence factor is above a threshold at 412 and if the answer is no, ….)
causing display of a control at a user interface, wherein the control is selectable to indicate a true authenticity of the item … (regarding causing display of a control at the user interface, wherein the control is selectable to indicate a true authenticity of the item: 0072 UI is presented for providing inputs, including designating card (0001, 0003, 0005 e.g., game card, whose authenticity is being evaluated) as authentic; regarding to indicate a true authenticity of the item: per 0100/Fig. 5, the 'designating as authentic' taught by 0072 occurs in a context where the human grader can designate an item as authentic (a first value) or counterfeit (a second value) ("confirmed (a first value) or rejected (a second value) by … human graders"))
based on the confidence score failing to satisfy the at least one threshold value; and (0119, Fig. 8, decision block 412, "No" branch; "Back at 412, the process queries whether the confidence factor is above a threshold at 412 and if the answer is no, …. If the analyzed card was not previously graded, it can be graded as discussed elsewhere [i.e., in the descriptions of Fig. 5 and 6] herein." -- that is to say, the grading of Fig. 5 or 6, which teaches to indicate a first value or a second value associated with a true authenticity of the item, is performed following (based on) the process of Fig. 8, in which it is determined that the confidence score fails to satisfy a threshold value)
receiving a selection via the control. (regarding via the control: 0072; regarding receiving a selection: 0100 "confirmed (receiving a selection of a first value) or rejected (receiving a selection of a second value) by … human graders")
It would have been obvious to one of ordinary skill in the art not later than the effective filing date of the claimed invention to have modified the combination of Ramasubramanian's systems and methods for identifying counterfeits, as modified by Asendorf's teachings regarding dividing an image into patches (segments) based on characteristic(s) of the image, by incorporating therein these teachings of Frisbee regarding performing a human review of an ML model's output of a grade/evaluation of an object for authenticity, following (based on) a confidence value failing to satisfy a threshold value, and receiving input (human grading/evaluation) via a UI, because it would increase accuracy of the output, see Frisbee, 0073, and can be used to improve the ML model by feeding back the human-determined result to retrain the model, see Frisbee, 0070.
Claims 9, 10 and 16-18 are rejected under 35 U.S.C. 103 as being unpatentable over Ramasubramanian et al. (U.S. Patent Application Publication No. 2022/0351215 A1), hereafter Ramasubramanian, in view of Asendorf et al. (U.S. Patent Application Publication No. 2020/0364513 A1), hereafter Asendorf, further in view of Frisbee et al. (U.S. Patent Application Publication No. 2022/0036371 A1), hereafter Frisbee, and further in view of Bean et al. (U.S. Patent Application Publication No. 2022/0353273 A1), hereafter Bean.
Regarding Claim 9
Ramasubramanian in view of Asendorf and Frisbee teaches the limitations of base claim 1 and intervening claim 8 as set forth above. Frisbee further teaches:
receiving a selection of the first value via the control; and (regarding via the control: 0072; regarding receiving a selection of the first value: 0100 "confirmed (receiving a selection of the first value) or rejected (receiving a selection of the second value) by … human graders")
It would have been obvious to one of ordinary skill in the art not later than the effective filing date of the claimed invention to have modified the combination of Ramasubramanian's systems and methods for identifying counterfeits, as modified by Asendorf's teachings regarding dividing an image into patches (segments) based on characteristic(s) of the image, and as further modified by Frisbee's teachings of performing a human review of an ML model's output of a grade/evaluation of an object for authenticity, following (based on) a confidence value failing to satisfy a threshold value, by incorporating therein these further teachings of Frisbee regarding receiving input (human grading/evaluation) via a UI, because the combination heretofore is merely potential ("the control is selectable to indicate") and these further teachings provide for the actualization of the potential ("receiving a selection") so as to render the combination working in practice, and because the combination is a matter of combining prior art elements according to known methods to yield predictable results. MPEP 2143.I.A.
Ramasubramanian in view of Asendorf and Frisbee does not explicitly disclose but Bean teaches:
processing a data transaction associated with the item based on the selection. (0041, 0044; see 0039-0040, 0042-0043 for context)
It would have been obvious to one of ordinary skill in the art not later than the effective filing date of the claimed invention to have modified the combination of Ramasubramanian's systems and methods for identifying counterfeits, as modified by Asendorf's teachings regarding dividing an image into patches (segments) based on characteristic(s) of the image, and as further modified by Frisbee's teachings of performing a human review of an ML model's output of a grade/evaluation of an object for authenticity, following (based on) a confidence value failing to satisfy a threshold value, and receiving input (human grading/evaluation) via a UI, by incorporating therein these teachings of Bean regarding preventing or permitting transactions depending on whether the object of the transaction has been determined counterfeit or authentic, respectively, because it would permit legitimate transactions and prevent illegitimate transactions.
Regarding Claim 10
Ramasubramanian in view of Asendorf and Frisbee teaches the limitations of base claim 1 and intervening claim 8 as set forth above. Frisbee further teaches:
receiving a selection of the second value via the control; and (regarding via the control: 0072; regarding receiving a selection of the second value: 0100 "confirmed (receiving a selection of the first value) or rejected (receiving a selection of the second value) by … human graders")
It would have been obvious to one of ordinary skill in the art not later than the effective filing date of the claimed invention to have modified the combination of Ramasubramanian's systems and methods for identifying counterfeits, as modified by Asendorf's teachings regarding dividing an image into patches (segments) based on characteristic(s) of the image, and as further modified by Frisbee's teachings of performing a human review of an ML model's output of a grade/evaluation of an object for authenticity, following (based on) a confidence value failing to satisfy a threshold value, by incorporating therein these further teachings of Frisbee regarding receiving input (human grading/evaluation) via a UI, because the combination heretofore is merely potential ("the control is selectable to indicate") and these further teachings provide for the actualization of the potential ("receiving a selection") so as to render the combination working in practice, and because the combination is a matter of combining prior art elements according to known methods to yield predictable results. MPEP 2143.I.A.
Ramasubramanian in view of Asendorf and Frisbee does not explicitly disclose but Bean teaches:
canceling processing of a data transaction associated with the item based on the selection. (0041, 0044; see 0039-0040, 0042-0043 for context)
It would have been obvious to one of ordinary skill in the art not later than the effective filing date of the claimed invention to have modified the combination of Ramasubramanian's systems and methods for identifying counterfeits, as modified by Asendorf's teachings regarding dividing an image into patches (segments) based on characteristic(s) of the image, and as further modified by Frisbee's teachings of performing a human review of an ML model's output of a grade/evaluation of an object for authenticity, following (based on) a confidence value failing to satisfy a threshold value, and receiving input (human grading/evaluation) via a UI, by incorporating therein these teachings of Bean regarding preventing or permitting transactions depending on whether the object of the transaction has been determined counterfeit or authentic, respectively, because it would permit legitimate transactions and prevent illegitimate transactions.
Regarding Claim 16
Ramasubramanian in view of Asendorf and Frisbee teaches the limitations of base claim 12 and intervening claim 15 as set forth above. Ramasubramanian in view of Asendorf and Frisbee does not explicitly disclose but Bean teaches:
wherein processing or canceling the one or more data transactions associated with the item comprises processing the one or more data transactions based on the selection indicating that the true authenticity of the item is authentic. (0041, 0044; see 0039-0040, 0042-0043 for context)
NOTE: for the rejection of claim 16, Bean instead of Ramasubramanian is cited as teaching the following limitation of claim 12:
processing or canceling, by the computing device, one or more data transactions associated with the item based on the binary value ("authenticity indicator") that indicates the authenticity of the item and the confidence score ("confidence level") associated with the authenticity of the item. (0039-0044)
It would have been obvious to one of ordinary skill in the art not later than the effective filing date of the claimed invention to have modified the combination of Ramasubramanian's systems and methods for identifying counterfeits, as modified by Asendorf's teachings regarding dividing an image into patches (segments) based on characteristic(s) of the image, and as further modified by Frisbee's teachings of performing a human review of an ML model's output of a grade/evaluation of an object for authenticity, following (based on) a confidence value failing to satisfy a threshold value, and receiving input (human grading/evaluation) via a UI, by incorporating therein these teachings of Bean regarding preventing or permitting transactions depending on whether the object of the transaction has been determined counterfeit or authentic, respectively, because it would permit legitimate transactions and prevent illegitimate transactions.
Regarding Claim 17
Ramasubramanian in view of Asendorf and Frisbee teaches the limitations of base claim 12 and intervening claim 15 as set forth above. Ramasubramanian in view of Asendorf and Frisbee does not explicitly disclose but Bean teaches:
wherein processing or canceling the one or more data transactions associated with the item comprises canceling the one or more data transactions based on the selection indicating that the true authenticity of the item is counterfeit. (0041, 0044; see 0039-0040, 0042-0043 for context)
It would have been obvious to one of ordinary skill in the art not later than the effective filing date of the claimed invention to have modified the combination of Ramasubramanian's systems and methods for identifying counterfeits, as modified by Asendorf's teachings regarding dividing an image into patches (segments) based on characteristic(s) of the image, and as further modified by Frisbee's teachings of performing a human review of an ML model's output of a grade/evaluation of an object for authenticity, following (based on) a confidence value failing to satisfy a threshold value, and receiving input (human grading/evaluation) via a UI, by incorporating therein these teachings of Bean regarding preventing or permitting transactions depending on whether the object of the transaction has been determined counterfeit or authentic, respectively, because it would permit legitimate transactions and prevent illegitimate transactions.
Regarding Claim 18
Ramasubramanian in view of Asendorf and Frisbee teaches the limitations of base claim 12 and intervening claim 15 as set forth above. Frisbee further teaches:
retraining the learning model based on the selection and the one or more image segments. (regarding retraining the learning model based on the selection …: 0070-0071, 0104, 0109; regarding retraining the learning model based on … the one or more image segments: 0070-0071, 0098, 0104, 0109 note the data used for retraining (0070, 0103-0104, 0108-0109) includes image data (e.g., 0101-0103), including image segments, e.g., 0102 ("In addition to measurements, the protocol can evaluate other card parameters and factors, including edges, corners, color or colors [image segments], …", and 0098 ("For example, the AI model can be re-trained or fine-tuned to improve detection of acceptable card edges. Finetuning can involve updating the CNN architecture and re-training it to learn new or different features of the different classes or different characteristics of the cards [image segments], such as different thresholds for acceptable card edges [image segments]."))
It would have been obvious to one of ordinary skill in the art not later than the effective filing date of the claimed invention to have modified the combination of Ramasubramanian's systems and methods for identifying counterfeits, as modified by Asendorf's teachings regarding dividing an image into patches (segments) based on characteristic(s) of the image, and as further modified by Frisbee's teachings of performing a human review of an ML model's output of a grade/evaluation of an object for authenticity, following (based on) a confidence value failing to satisfy a threshold value, and receiving input (human grading/evaluation) via a UI, by incorporating therein these further teachings of Frisbee regarding retraining the machine learning model, because retraining is a standard, fundamental aspect of machine learning/machine learning training, which serves to improve accuracy of the model and hence of the model output/results, see Frisbee, 0070-0071, 0098, and as such the combination is merely a matter of combining prior art elements according to known methods to yield predictable results; use of known technique to improve similar devices (methods, or products) in the same way; and applying a known technique to a known device (method, or product) ready for improvement to yield predictable results. MPEP 2143.I.A.,C.,D.
Conclusion
The prior art made of record and not relied upon, as set forth in the accompanying Notice of References Cited (PTO-892), is considered pertinent to applicant's disclosure. Among the cited documents:
Sharma (US-20210117984-A1) teaches a method and system to determine whether a suspect item is counterfeit, including analyzing images in order to identify regions of interest (ROI) of the item, comparing characteristics of different ROIs against threshold characteristics derived from an item model, and determining whether the suspect item is counterfeit based on a result of the comparison, and more specifically (see Fig. 2B) including segmenting an image into ROIs using a neural network, constructing a classification vector based on ROI vectors, inputting the classification vector into a machine learning model, and outputting from the machine learning model an authenticity decision about the item.
Ives (US-20250217817-A1) teaches a system and method for detecting counterfeit items, that allows for a consumer to verify the authenticity of a commercial product at the point of sale using his or her mobile phone, wherein images of the product packaging, captured by the user and sent to a remote server, are analyzed at the server and the consumer receives a result from the server saying whether the product is a genuine product and/or whether the product is being sold legally, and further including determining a risk profile for an item, and a confidence score based at least on the risk profile, the confidence score giving an indication of how likely it is that the item is an authentic item, including use of a machine-readable mark encoding a unique identifier of the item in order to determine legitimacy of the item.
Lei (US-20090152357-A1) teaches identifying and authenticating/validating a security document (e.g., driver's license) including inter alia (see Figs. 10, 12) capturing image data, segmenting the image, calculating weight coefficient vectors, comparing the document image to a reference document and determining if the document is of the document type of the reference document.
Sundararaman (US-20210004580-A1) teaches transaction auditing including training a machine learning model to determine features that can be used to determine whether an image is an authentic image of a document (e.g., receipt) or an automatically generated document image, using a training data, wherein the machine learning model classifies the image as either an authentic image of a document or an automatically generated document image, based on features included in the image that are identified by the machine learning model, and the machine learning model is updated based on the image and the classification of the image.
Callegari (US-20160300107-A1) teaches image analysis for authenticating a product (e.g., a wristwatch), including taking an image of the product and comparing the image with a reference image of a genuine product taken previously to determine if the products in the two images are the same, wherein the two images are processed in order to calculate for each of them a list of significant points, the significant points are compared (in a common coordinate system defined by some of the significant points that match) to determine a degree of correspondence between the significant points, and an answer is output indicating the authenticity of the product based on the degree of correspondence.
Piramuthu (US-20190205962-A1) teaches a system for computer vision and image characteristic search, including determining visual characteristics of objects depicted in images and comparing the determined characteristics to visual characteristics of other images, e.g., to identify similar visual characteristics in the other images, and including pattern-based authentication in which the system determines authenticity of an item in an image based on a similarity of its visual characteristics to visual characteristics of known authentic items, and including inter alia (Fig. 12) allowing / not allowing an item to be listed for sale based on whether it is determined to be authentic or not.
Burgin (US-20190236614-A1) teaches an AI counterfeit detection system that implements a Generative Adversarial Network (GAN) to classify an image as one of a fake or genuine item and integrates a Classification Activation Module (CAM) to refine counterfeit detection, wherein the GAN includes a generator that generates simulated counterfeit images for a discriminator, and the discriminator is trained to identify faked items by learning from the simulated counterfeit images and/or images of actual faked items, the discriminator may implement a deep neural network of convolutional layers that each analyze a region of an image and produce a weighted output that contributes to the classification based on the analyzed region, and the CAM may identify the regions and weights relied upon by the discriminator, provide corresponding heatmaps to subject matter experts, receive annotations from the subject matter experts, and use the annotations as feedback to refine the classifier of the discriminator.
Chaloux (US-20190354744-A1) teaches detection of counterfeit items based on machine learning and analysis of visual and textual data, including receiving multimodal product data, searching terms based on the received product data, identifying potential counterfeit items a web crawling technique, identifying matches for similar products, using image processing and analysis to determine if the matches comprise potential counterfeit items, and generating a takedown notice if a user confirms the items are counterfeits.
Oda (US-20250111685-A1) (qualifying as prior art only based on priority document) teaches methods and apparatus to analyze an image of a portion of an item for a pattern indicating authenticity of the item, including receiving a plurality of images having a plurality of image types; for each image type from the plurality of image types and to generate a plurality of subsets of images, identifying a subset of images from the plurality of images being that image type using an image classifier; for each subset of images from the plurality of subsets of images, performing feature extraction on each image from that subset of images to generate features associated with that image; inputting the features associated with each image from that subset of images to a trained ML model from a plurality of trained ML models to generate an output indicating whether a collectible associated with that image is authentic or counterfeit.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any extension fee pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DOUGLAS W PINSKY whose telephone number is (571)272-4131. The examiner can normally be reached on 8:30 am - 5:30 pm ET.
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, Jessica Lemieux can be reached on 571-270-3445. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/DOUGLAS W PINSKY/
Examiner, Art Unit 3626
/JESSICA LEMIEUX/Supervisory Patent Examiner, Art Unit 3626
1 Note as per the specification (0001, 0058) "learning" refers to "machine learning" or "artificial intelligence."
2 This limitation is the same as the corresponding limitation in claims 1 and 11; the prior art cited for this limitation here in claim 12 is the same as is cited for the corresponding limitation in claims 1 and 11.
3 This limitation is the same as the corresponding limitation in claims 1 and 11; the prior art cited for this limitation here in claim 12 is the same as is cited for the corresponding limitation in claims 1 and 11.
4 This limitation is the same as the corresponding limitation in claims 1 and 11; the prior art cited for this limitation here in claim 12 is the same as is cited for the corresponding limitation in claims 1 and 11.
5 This limitation is the same as the corresponding limitation in claims 1 and 11; the prior art cited for this limitation here in claim 12 is the same as is cited for the corresponding limitation in claims 1 and 11.
6 Note alternatively Cella also teaches:
wherein the confidence score fails to satisfy a threshold value, the computer-implemented method further comprising: (0709 "training data samples that appear to be difficult to classify correctly and/or with high confidence) may be submitted to a human reviewer, and the semi-supervised learning process may receive, from the human reviewer, one or more labels that correspond to an expected and/or desirable output of the graph neural network for such portions of the input training data"; note the term "high" is a relative term, hence necessarily relative to some baseline/ benchmark, i.e., threshold)
receiving, … an indication of the authenticity of the item; and (0709 "training data samples that appear to be difficult to classify correctly and/or with high confidence) may be submitted to a human reviewer, and the semi-supervised learning process may receive, from the human reviewer, one or more labels that correspond to an expected and/or desirable output of the graph neural network for such portions of the input training data")
7 Note alternatively Frisbee also teaches:
wherein the confidence score fails to satisfy a threshold value, the computer-implemented method further comprising: (0119, Fig. 8, decision block 412, "No" branch)
Prosecution Insights