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 .
Claim Status
The status of claims 1-20 is:
Claims 1-20 are pending.
Claim Objections
Claims 10 and 20 are objected to because of the following informalities: the claim states “on validation” when it should be “one validation”. Appropriate correction is required.
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claim(s) 1-2, 4-6, 8-12, 14-16, and 18-20 are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Matias et al. (U.S. Patent Publication No 2025/0148785, hereinafter “Matias”).
Regarding claim 1, Matias discloses a method for analyzing image data, the method comprising:
detecting, by a device, image data (Matias [0005]: “According to a first aspect, a computer implemented method of detecting a deepfake video, comprises: feeding a video into at least one forensic model”);
controlling, by the device, at least one analysis operation to analyze the image data (Matias [0043]: “An input video is fed into one or more forensic models. An indication(s) of one or more properties of a person and/or object depicted in the video is obtained as an outcome of the forensic model(s). For example, height, weight, size, medical condition, native tongue, and tiredness (when available as ground truth)”), wherein the at least one analysis operation includes an operation to perform object recognition of a first object in the image data (Matias [0043]: “An input video is fed into one or more forensic models. An indication(s) of one or more properties of a person and/or object depicted in the video is obtained as an outcome of the forensic model(s). For example, height, weight, size, medical condition, native tongue, and tiredness (when available as ground truth)”, implies object recognition since it would be required in order to determine a property/characteristic) and an operation to determine a characteristic of the first object using the image data (Matias [0043]: “An input video is fed into one or more forensic models. An indication(s) of one or more properties of a person and/or object depicted in the video is obtained as an outcome of the forensic model(s). For example, height, weight, size, medical condition, native tongue, and tiredness (when available as ground truth)”);
controlling, by the device, at least one validation operation to verify the characteristic of the first object, the validation operation generating at least one validation output (Matias [0043]: “The indication(s) is compared to a ground truth(s) obtained by applying the forensic model(s) to an authentic video(s) depicting the same or similar person and/or object, and/or by searching a ground truth dataset storing previously generated ground truths which are known to be authentic. Likelihood that the video is deepfake is increased in response to a mismatch between the indication(s) and the ground truth(s). Alternatively, likelihood that the video is authentic, or at least the current part being analyze is authentic, is increased in response to a match between the indication(s) and the ground truth(s)”); and
controlling, by the device, output of the at least one validation output, wherein the output of the at least one validation operation includes output of at least one graphical element with the image data (Matias [0102]: “At 216, one or more actions may be triggered in response to the likelihood (e.g., probability that the input video is a deepfake video. For example, the likelihood is presented on a display, viewing of the input video is automatically blocked, the input video is automatically deleted, the input video is added to a dataset of known deepfake videos to enable quick detection of the same or similar deepfake video (e.g., without running through the rest of the features of FIG. 2), and/or a message is automatically sent to another server for example, an administrative server. In another example, a report may be generated. The report may be generated in response to a match and/or mismatch of comparing an outcome of the forensic model(s) to the ground truth, for example, as described with reference to FIG. 4. The report may indicate one or more of: the match or mismatch, a value of the increase in likelihood that the video is deepfake or authentic, and in response to a mismatch providing the at least one indication outcome of the at least one forensic model and the at least one ground truth. For example, when using a forensic model that estimates age in response to an input of spoken speech of a person, the report may indicate: outcome reduced likelihood of authenticity by 25-40%, the ground truth age of the person is 25 years while the output of the forensic model indicates that the age of the person is 60-75 years”).
Regarding claim 11, it is rejected under the same analysis as claim 1 above along with Matias’ disclosure of an interface (Matias [0073]: “Computing device 104 and/or client terminal(s) 108 include and/or are in communication with one or more physical user interfaces 126 that include a mechanism for a user to enter data (e.g., manually designate the location of video 150 for analysis) and/or view the displayed results (e.g., indication of whether video 150 is a deepfake video and/or specific deepfake tool used to create video 150), within a GUI”) a memory storing executable instructions (Matias [0024]: “A non-transitory medium storing program instructions for detecting a deepfake video, which when executed by at least one processor, cause the at least one processor to”); and a controller (Matias [0024]: “A non-transitory medium storing program instructions for detecting a deepfake video, which when executed by at least one processor, cause the at least one processor to”).
Regarding claim 2, Matias discloses the method, wherein detecting image data includes detecting image data in at least one of a data stream, display content and electronic file (Matias [0043]: “An input video is fed into one or more forensic models. An indication(s) of one or more properties of a person and/or object depicted in the video is obtained as an outcome of the forensic model(s)”, a video is a type of data stream), and wherein the image data is received by the device (Matias [0043]: “An input video is fed into one or more forensic models. An indication(s) of one or more properties of a person and/or object depicted in the video is obtained as an outcome of the forensic model(s)”, a video is a type of data stream).
Regarding claim 12, it is rejected under the same analysis as claim 2 above.
Regarding claim 4, Matias discloses the method, wherein the at least one analysis operation to analyze image data includes detecting a two-dimensional characteristic of the first object in the image data (Matias [0088]: “For example, a forensic model that measures height from voice, computes a height of 176 centimeters (cm) from an input voice of a certain person depicted in an input video”), and wherein the validation operation includes verifying the characteristic of the first object to at least one parameter characteristic for the first object (Matias [0088]: “When an authentic video of the certain person is analyzed by the forensic model, and is determined that the height of the certain person is 178 cm, the property may be validated, increasing likelihood that the input video is authentic. In contrast, when an analysis of the authentic video of the certain person by the forensic model indicates that the height of the certain person is 250 cm, a discrepancy is detected, increasing likelihood that the voice is deepfake”).
Regarding claim 14, it is rejected under the same analysis as claim 4 above.
Regarding claim 5, Matias discloses the method, wherein the at least one analysis operation to analyze image data includes detecting an inconsistency of the object in the image data (Matias [0088]: “When an authentic video of the certain person is analyzed by the forensic model, and is determined that the height of the certain person is 178 cm, the property may be validated, increasing likelihood that the input video is authentic. In contrast, when an analysis of the authentic video of the certain person by the forensic model indicates that the height of the certain person is 250 cm, a discrepancy is detected, increasing likelihood that the voice is deepfake”).
Regarding claim 15, it is rejected under the same analysis as claim 5 above.
Regarding claim 6, Matias discloses the method, wherein the at least one analysis operation to analyze image data includes detecting use of preexisting image data (Matias [0088]: “When an authentic video of the certain person is analyzed by the forensic model, and is determined that the height of the certain person is 178 cm, the property may be validated, increasing likelihood that the input video is authentic. In contrast, when an analysis of the authentic video of the certain person by the forensic model indicates that the height of the certain person is 250 cm, a discrepancy is detected, increasing likelihood that the voice is deepfake”).
Regarding claim 16, it is rejected under the same analysis as claim 6 above.
Regarding claim 8, Matias discloses the method, wherein the at least one analysis operation to analyze image data includes detecting a machine model generated element in the image data (Matias [0005]: “According to a first aspect, a computer implemented method of detecting a deepfake video”).
Regarding claim 18, it is rejected under the same analysis as claim 8 above.
Regarding claim 9, Matias discloses the method, wherein the graphical element with the image data identifies at least one characteristic of the first object (Matias [0102]: “In another example, a report may be generated. The report may be generated in response to a match and/or mismatch of comparing an outcome of the forensic model(s) to the ground truth, for example, as described with reference to FIG. 4. The report may indicate one or more of: the match or mismatch, a value of the increase in likelihood that the video is deepfake or authentic, and in response to a mismatch providing the at least one indication outcome of the at least one forensic model and the at least one ground truth. For example, when using a forensic model that estimates age in response to an input of spoken speech of a person, the report may indicate: outcome reduced likelihood of authenticity by 25-40%, the ground truth age of the person is 25 years while the output of the forensic model indicates that the age of the person is 60-75 years”).
Regarding claim 19, it is rejected under the same analysis as claim 9 above.
Regarding claim 10, Matias discloses the method, further comprising controlling at least one machine model to generate reference image data using the object recognition of the first object as input, and wherein controlling at least one validation operation includes comparison of the reference image data to the image data (Matias [0088]: “For example, a forensic model that measures height from voice, computes a height of 176 centimeters (cm) from an input voice of a certain person depicted in an input video. When an authentic video of the certain person is analyzed by the forensic model, and is determined that the height of the certain person is 178 cm, the property may be validated, increasing likelihood that the input video is authentic. In contrast, when an analysis of the authentic video of the certain person by the forensic model indicates that the height of the certain person is 250 cm, a discrepancy is detected, increasing likelihood that the voice is deepfake”).
Regarding claim 20, it is rejected under the same analysis as claim 10 above.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 3 and 13 are rejected under 35 U.S.C. 103 as being unpatentable over Matias in view of Ye et al. (U.S. Patent Publication No 2021/0117650, hereinafter “Ye”).
Regarding claim 3, Matias does not explicitly disclose the method, wherein the at least one analysis operation to analyze image data includes detecting a lighting characteristic of the first object relative to at least one image area of the image data.
However, Ye teaches the method, wherein the at least one analysis operation to analyze image data includes detecting a lighting characteristic of the first object relative to at least one image area of the image data (Ye [0008]: “In another aspect, a method includes processing an image through a face detection module to output feature vectors indicating at least one lighting irregularity on a face in the image, or at least one texture irregularity in the image, or both. The method also includes processing the image through at least one discrete Fourier transform (DFT) and at least one neural network to output feature vectors indicating at least one irregularity in the image in the frequency domain, and returning an indication that the image has been altered from an original image at least in part based on the feature vectors”; Ye [0092]: “FIGS. 14-16 provide additional examples of artifacts or irregularities that can appear in altered images, labeled “fake” images in the figures. A first real image 1400 in FIG. 14 has been altered to produce a corresponding altered image 1402 in which, in a region 1404, lighting appears to be brighter than in the corresponding region in the first real image 1400. Likewise, a second real image 1406 has been altered to produce an altered image 1408 in which lighting in a region 1410 on the face appears to be brighter than in the real image 1406. The resolutions of the altered images 1402, 14108 are also less than the resolutions of the corresponding real image 1400, 1406, meaning that the NNs can learn to distinguish altered images on the basis of either or both lighting irregularities and resolution decreases”).
It would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to incorporate detecting lighting irregularities as taught by Ye with the method of Matias because it would increase the number of features that the method is able to evaluate which will make it better able to detect generated images (Ye [0092). This motivation for the combination of Matias and Ye is supported by KSR exemplary rationale (G) Some teaching, suggestion, or motivation in the prior art that would have led one of ordinary skill to modify the prior art reference or to combine prior art reference teachings to arrive at the claimed invention and rationale (D) Applying a known technique to a known device (method, or product) ready for improvement to yield predictable results.
Regarding claim 13, it is rejected under the same analysis as claim 3 above.
Claim(s) 7 and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Matias in view of Xu et al. (Xu, Z., Zhang, X., Li, R., Tang, Z., Huang, Q., & Zhang, J. (2024). FAKESHIELD: EXPLAINABLE IMAGE FORGERY DE. arXiv preprint arXiv:2410.02761., hereinafter “Xu”).
Regarding claim 7, Matias does not explicitly disclose the method, wherein the at least one analysis operation to analyze image data includes detecting pixel blur in the image data.
However, Xu teaches the method, wherein the at least one analysis operation to analyze image data includes detecting pixel blur in the image data (Xu Page 17: “(4) Resolution: Please check the resolution and compression traces of the picture. Composite or edited images may show unnatural pixel blur, jaggedness, or excessive compression”).
It would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to incorporate detecting pixel blur as taught by Xu with the method of Matias because edited images can show pixel blur and so being able to detect pixel blur would improve the ability of the method to detect generated images (Xu Page 17). This motivation for the combination of Matias and Xu is supported by KSR exemplary rationale (G) Some teaching, suggestion, or motivation in the prior art that would have led one of ordinary skill to modify the prior art reference or to combine prior art reference teachings to arrive at the claimed invention and rationale (D) Applying a known technique to a known device (method, or product) ready for improvement to yield predictable results.
Regarding claim 17, it is rejected under the same analysis as claim 7 above.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to AIDAN KEUP whose telephone number is (703)756-4578. The examiner can normally be reached Monday - Friday 8:00-4:00.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Emily Terrell can be reached at (571) 270-3717. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/AIDAN KEUP/ Examiner, Art Unit 2666 /Molly Wilburn/Primary Examiner, Art Unit 2666
Prosecution Insights