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 .
Notice to Applicants
This action is in response to the Application filed on 11/21/2024.
Claims 1-20 are pending.
Information Disclosure Statement
The Information Disclosure Statement (IDS) filed on 11/21/2024 has been fully considered by the examiner.
Claim Objections
Claims 8 and 13-20 are objected to.
Regarding claim 8, lines 2-4 should be amended to read “the image generation model is trained to generate images at a first scale factor corresponding to the global context and at a second scale factor corresponding to the local context.” (emphasis added).
Regarding claim 13, line 2 should be amended to read “the intermediate inpainting result is generated during a first denoising phase” (emphasis added).
Regarding claim 14, the same objection applies to claim 14 based on its dependence on claim 13.
Regarding claim 15, lines 3-5 should be amended to read “wherein the processing device is configured to execute instructions stored in the memory component to perform operations comprising :” (emphasis added).
Regarding claims 16-20, the same objection applies to claims 16-20 based on their dependence on claim 15.
Claim Rejections – 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 4-8, 10, and 18 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Regarding claims 4, 6-8, 10, and 18, these claims each recite terms that lack antecedent basis in their respective claim:
Claim 4: “the first scale” and “the second scale”,
Claim 6: “the first denoising phase”,
Claims 7-8 inherit this issue,
Claim 7: “the second denoising phase”,
Claim 10: “the first scale factor” and “the second scale factor”,
Claim 18: “the first scale” and “the second scale”.
Regarding claim 5, the claim recites that the intermediate inpainting result “depicts the scene with synthetic content consistent with the global context in the region indicated by the inpainting indication” (emphasis added). The examiner argues that “consistency with the global context” is a subjective term that the specification does not provide a comprehensive way to measure as claimed.
The broadest reasonable interpretation of “global context” as claimed includes the entirety of an image containing an inpainted region, or a subset of said image, along with the semantic features included therein.
The originally-filed specification does provide both examples of synthetic content that are inconsistent and consistent with the global context, such as inpainted body parts being of a different/same skin color or lighting matching/not matching the rest of the scene (see figures 2-3 and paragraphs 0031-0033). However, these examples are non-limiting, and as stated by MPEP 2173.05(b).IV, "[f]or some facially subjective terms, the definiteness requirement is not satisfied by merely offering examples that satisfy the term within the specification."
Although the provided examples in the specification are clear, the claim encompasses all possible examples of inpainted content being consistent with the global context of any input image. This raises a substantial amount of scenarios where a subjective determination of “consistency” would be required.
As just one example, imagine an image of a paper with a handwritten sentence, and a word of said sentence is inpainted. As long as the inpainted word makes the sentence grammatically correct, then it can be argued to be consistent. However, one can also consider if the inpainted word matches the handwriting of the other words in the image, and this question would have different, subjective answers depending on the person, especially if the inpainted word does not appear anywhere else in the original image to compare to.
In summary, the scope of the claim encompasses every imaginable scenario of inpainted content being consistent with other parts of the image, and without any definitive way to determine said consistency, there are many cases where subjective judgement would be required to measure “consistency”, and thus the claim is indefinite.
The examiner notes that claim 5 recites “wherein the inpainted image includes the synthetic content with a higher level of detail than in the intermediate inpainting result”, and claim 9 (and claims 10-14 via dependency on claim 9) recite “wherein the inpainted image depicts the scene with the synthetic content in the region indicated by the inpainting indication at a higher level of detail than in the intermediate inpainting result” (emphasis added in both).
At face value, this also appears to be a subjective term, but the specification does provide a way for one of ordinary skill in the art to interpret “detail” objectively. Paragraph 0022 states that “the global crop and the local crop can be scaled differently so that they both match the target input size of the image generation model.” If it is assumed that both the claimed intermediate inpainting result and the inpainted image are both of the same resolution, i.e. the target input size, then either the global crop would need to be downscaled, or the local crop upscaled, or possibly both, to match the resolution. In either case, the inpainted region would be of a higher resolution in the local crop, and thus the synthetic content would be processed at a higher resolution, or at a higher level of detail.
35 USC § 101 Analysis
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.
The examiner determines that all of claims 1-20 are eligible under 35 U.S.C. 101. The 101 analysis is provided below for purposes of a clear record.
Analysis for claim 1 is provided in the following. Claim 1 is reproduced in the following (annotation added):
A method comprising:
obtaining an input image and an inpainting indication, wherein the input image depicts a scene and the inpainting indication indicates a region of the scene to be inpainted;
generating, using an image generation model, an intermediate inpainting result based on a global context of the input image;
and generating, using the image generation model, an inpainted image based on the intermediate inpainting result and a local context of the input image,
wherein the inpainted image depicts the scene with synthetic content in the region indicated by the inpainting indication.
Step 1: Does the claim belong to one of the statutory categories? Claim 1 is directed to a process, which is a statutory category of invention (YES).
Step 2A Prong One: Does the claim recite a judicial exception? The claim does not appear to recite any judicial exceptions. The recited steps of using the image generation model to produce inpainted images inherently involve mathematical calculations, but the claim does not set forth any specific mathematical calculations. The obtaining of the inpainting indication in step b could be considered a mental process, where a human mentally determines what section of the image should be inpainted, but in this case the remaining inpainting steps are not performable in the human mind, and would plainly integrate this mental process into a practical application (NO). Claim 1 is eligible.
Similar analysis is applicable to independent claim 9. Claim 9 recites a non-transitory computer readable medium storing code that, when executed by a processor, performs similar inpainting steps to claim 1. The main difference is that the inpainting steps here are performed at different scales, with the second scale being smaller than the first scale, and that the final inpainting result is of a higher level of detail than the intermediate result, none of which are directed to any judicial exceptions. Claim 9 is eligible.
Similar analysis is applicable to independent claim 15. Claim 15 recites a computerized system that performs similar inpainting steps to claim 1, with the main difference being that the local context “includes a smaller portion of the input image than the global context”, which is still not directed to any judicial exceptions. Claim 15 is eligible.
Claims 2 and 16 recite identifying a first scale factor and a second, smaller scale factor, which can be practically performed in the human mind. However, the claims further recite that the global/local contexts are based on the first/second factors, respectively, which integrates this mental process into the inpainting application. Claims 2 and 16 are eligible.
Claims 3 and 17 recite cropping the image based on the first/second scale factors to obtain the global/local contexts, respectively, which also integrates the mental process of claim 2 into a practical application. Claims 3 and 17 are eligible.
Claims 4-5, 8, 10-12, 18, and 20 do not recite any new judicial exceptions. Claims 4-5, 8, 10-12, 18, and 20 are eligible.
Claims 6, 13, and 19 recite identifying an intermediate diffusion timestep, which can be practically performed in the human mind. However, each of these claims also recite that the first denoising phase includes denoising iterations performed prior to this identified timestep, and thus this mental process is integrated into the practical denoising application. Claims 6, 13, and 19 are eligible.
Claims 7 and 14 recite that the second denoising phase includes denoising iterations performed after the identified timestep, which also integrates the mental process of claims 6/13, respectively, into the practical denoising application. Claims 7 and 14 are eligible.
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.
Claims 1, 4, 9, 11-12, 15, and 18 are rejected under 35 U.S.C. 102(a)(1) and 102(a)(2) as being anticipated by Kim et al. (U.S. Publ. US-2023/0325985-A1).
Regarding claim 1, Kim discloses a method (see figure 2A and paragraphs 0043-0048) comprising:
obtaining an input image and an inpainting indication, wherein the input image depicts a scene and the inpainting indication indicates a region of the scene to be inpainted (see figure 2B, input image 202, subject 202A, masked regions 202B and paragraphs 0049-0051, where the input image depicts a scene including subject 202A and includes masked regions 202B to be inpainted);
generating, using an image generation model, an intermediate inpainting result based on a global context of the input image (see figure 2B, first inpainted image 206 and paragraphs 0049-0051, where the coarse network analyzes the entire image to generate the first inpainted image 206 with inpainted masked regions 206B);
and generating, using the image generation model, an inpainted image based on the intermediate inpainting result and a local context of the input image (first see figure 2C and paragraphs 0052-0053, where the first inpainted image and mask are input to a super-resolution network to produce a second inpainted image of a larger size/resolution; then see figure 2D and paragraphs 0054-0055, where this image is then input to a refinement network to produce accurate refinements to the inpainted regions; paragraphs 0038-0040 and 0047 specify that the refinement network focuses on local portions/contexts around the inpainted regions to improve finer details thereof),
wherein the inpainted image depicts the scene with synthetic content in the region indicated by the inpainting indication (as the refinement network in paragraphs 0054-0055 refines local portions of a larger-resolution image, the refined inpainted regions depict the synthetic content with higher details/resolution).
Regarding claim 4, Kim discloses generating a first inpainting mask based on the inpainting indication and the first scale (see figure 2B, mask 214 and paragraphs 0049-0051, where the mask is applied to the original image at its initial scale/resolution),
wherein the intermediate inpainting result is generated based on the first inpainting mask (see figure 2B, first inpainted image 206 and paragraphs 0049-0051, where the first image is generated according to the mask 214);
and generating a second inpainting mask based on the inpainting indication and the second scale (see figure 2C, second mask 218 and paragraphs 0052-0053, where the mask's resolution is also increased to match the resolution/scale of the second inpainted image 208),
wherein the inpainted image is generated based on the second inpainting mask (see figure 2D, first refined inpainted image 212 and paragraphs 0054-0055, where the refined image is generated according to the second mask).
Regarding claim 9, Kim discloses a non-transitory computer readable medium storing code for image processing (see figure 1, memory 106 and paragraph 0027),
the code comprising instructions that, when executed by at least one processor, cause the at least one processor to perform operations (see figure 1, processors 104 and paragraphs 0026, 0028) comprising:
obtaining an input image and an inpainting indication, wherein the input image depicts a scene and the inpainting indication indicates a region of the scene to be inpainted (see citations to same limitation in claim 1 above);
generating, using an image generation model, an intermediate inpainting result at a first scale (see figure 2B, first inpainted image 206 and paragraphs 0049-0051, where the coarse network analyzes the entire image to generate the first inpainted image 206 with inpainted masked regions 206B; the input image necessarily has a first scale/size/resolution);
and generating, using the image generation model, an inpainted image at a second scale that is smaller than the first scale based on the intermediate inpainting result (first see figure 2C and paragraphs 0052-0053, where the first inpainted image and mask are input to a super-resolution network to produce a second inpainted image of a larger size/resolution; then see figure 2D and paragraphs 0054-0055, where this image is then input to a refinement network to produce accurate refinements to the inpainted regions; paragraphs 0038-0040 and 0047 specify that the refinement network focuses on local portions/scales around the inpainted regions to improve finer details thereof),
wherein the inpainted image depicts the scene with the synthetic content in the region indicated by the inpainting indication at a higher level of detail than in the intermediate inpainting result (as the refinement network in paragraphs 0054-0055 refines local portions of a larger-resolution image, the refined inpainted regions depict the synthetic content with higher details/resolution).
Regarding claim 11, Kim discloses claim 11 as applied to claim 4 above.
Regarding claim 12, Kim discloses the intermediate inpainting result is generated based on a global context (see figure 2B, first inpainted image 206 and paragraphs 0049-0051, where the coarse network analyzes the entire image / global context to generate the first inpainted image)
and the inpainted image is generated based on local context (paragraphs 0038-0040 and 0047 specify that the refinement network focuses on local portions/contexts around the inpainted regions to improve finer details thereof),
and wherein the global context and the local context both comprise regions of the input image including the region of the scene to be inpainted (both of the above inpainting steps operate on regions of the image including the inpainting regions, i.e. the entire image and a portion of the image, respectively).
Regarding claim 15, Kim discloses a system (see figure 1, system 100 and paragraph 0025) comprising:
a memory component (see figure 1, memory 106 and paragraph 0027);
a processing device coupled to the memory component, wherein the processing device is configured to execute instructions stored in the memory component to perform operations (see figure 1, processors 104 and paragraphs 0026, 0028) comprising:
obtaining an input image and an inpainting indication, wherein the input image depicts a scene and the inpainting indication indicates a region of the scene to be inpainted (see citations to same limitation in claim 1 above);
generating, using an image generation model, an intermediate inpainting result based on a global context of the input image (see citations to same limitation in claim 1 above);
and generating, using the image generation model, an inpainted image based on the intermediate inpainting result and a local context of the input image that includes a smaller portion of the input image than the global context (first see figure 2C and paragraphs 0052-0053, where the first inpainted image and mask are input to a super-resolution network to produce a second inpainted image of a larger scale/size/resolution; then see figure 2D and paragraphs 0054-0055, where this image is then input to a refinement network to produce accurate refinements to the inpainted regions; paragraphs 0038-0040 and 0047 specify that the refinement network focuses on local portions/contexts around the inpainted regions to improve finer details thereof),
wherein the inpainted image depicts the scene with synthetic content in the region indicated by the inpainting indication (as the refinement network in paragraphs 0054-0055 refines local portions of a larger-resolution image, the refined inpainted regions depict the synthetic content with higher details/resolution).
Regarding claim 18, Kim discloses claim 18 as applied to claim 4 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.
Claims 2-3, 10, and 16-17 are rejected under 35 U.S.C. 103 as being unpatentable over Kim et al. (U.S. Publ. US-2023/0325985-A1) in view of Kudelski et al. (U.S. Publ. US-2022/0366544-A1).
The examiner notes for conciseness that, regarding claims 2-3, 10, and 16-17, although Kim resizes/rescales the images to different resolutions/scales, they merely fail to disclose using specific scale factors to determine the size of the inpainting regions, and they merely fail to crop the unprocessed part of the region out, such as the regions outside Kim's local portions.
Regarding claim 2, Kim fails to disclose the limitations of claim 2.
Pertaining to the same field of endeavor, Kudelski discloses identifying a first scale factor and a second scale factor that is smaller than the first scale factor, wherein the global context is based on the first scale factor and the local context is based on the second scale factor (see paragraphs 0043-0055, where an inpainting preprocessing step includes determining an "image crop" / global context that is any subset of the image, and a "mask crop" / local context focused on a smaller region around the inpainting mask; each image and mask crop is assigned a size/scale factor and can also be downscaled by a specific factor).
Kim and Kudelski are considered analogous art, as they are both directed to machine learning models for image inpainting. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have integrated the teachings of Kudelski into Kim by using scale factors to determine the global/local contexts because doing so allows for determining the optimal crop size to balance context with processing load (see Kudelski paragraphs 0046, 0050-0051).
Regarding claim 3, Kim fails to disclose the limitations of claim 3.
Pertaining to the same field of endeavor, Kudelski discloses cropping the input image based on the first scale factor to obtain the global context; and cropping the input image based on the second scale factor to obtain the local context (see paragraph 0053, where the image crop / global context and mask crop / local context are extracted from the image at their respective sizes/scales to be input to the inpainting model).
Kim and Kudelski are considered analogous art, as they are both directed to machine learning models for image inpainting. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have integrated the teachings of Kudelski into Kim by cropping the image to obtain the global/local contexts because doing so allows for determining the optimal crop size to balance context with processing load (see Kudelski paragraphs 0046, 0050-0051).
Regarding claim 10, Kim discloses
wherein the intermediate inpainting result is generated based on the global context (see figure 2B, first inpainted image 206 and paragraphs 0049-0051, where the coarse network analyzes the entire image to generate the first inpainted image 206 with inpainted masked regions 206B);
wherein the inpainted image is generated based on the local context (paragraphs 0038-0040 and 0047 specify that the refinement network focuses on local portions/contexts around the inpainted regions to improve finer details thereof).
Kim fails to disclose cropping according to the scale factors to obtain the global/local contexts, as indicated via strikethrough above.
Pertaining to the same field of endeavor, Kudelski discloses cropping the input image based on the first scale factor to obtain a global context, and cropping the input image based on the second scale factor to obtain a local context (see paragraph 0053, where the image crop / global context and mask crop / local context are extracted from the image at their respective sizes/scales to be input to the inpainting model).
Kim and Kudelski are considered analogous art, as they are both directed to machine learning models for image inpainting. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have integrated the teachings of Kudelski into Kim by cropping the image to obtain the global/local contexts because doing so allows for determining the optimal crop size to balance context with processing load (see Kudelski paragraphs 0046, 0050-0051).
Regarding claim 16, Kim in view of Kudelski discloses claim 16 as applied to claim 2 above.
Regarding claim 17, Kim in view of Kudelski discloses claim 17 as applied to claim 3 above.
Claims 5-7, 13-14, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Kim et al. (U.S. Publ. US-2023/0325985-A1) in view of Dey et al. (U.S. Publ. US-2026/0127730-A1).
Regarding claim 5, Kim discloses
and depicts the scene with synthetic content consistent with the global context in the region indicated by the inpainting indication (see paragraph 0035, where the coarse network extracts features for the entire image and inpaints the masked regions according to said features, thus the inpainted features can be said to be consistent with the global context / image);
and wherein the inpainted image includes the synthetic content with a higher level of detail than in the intermediate inpainting result (as the refinement network in paragraphs 0054-0055 refines local portions of a larger-resolution image, the refined inpainted regions depict the synthetic content with higher details/resolution).
Kim fails to disclose using a diffusion/denoising process to generate both the inpainting results, as indicated via strikethrough above.
Pertaining to the same field of endeavor, Dey discloses the intermediate inpainting result is generated by performing a first denoising process during a first denoising phase; and the inpainted image is generated by performing the denoising process during a second denoising phase (see paragraphs 0039-0051, 0058, where a DDPM / Denoising Diffusion Probabilistic Model is used to inpaint image regions via a denoising process; paragraph 0085 specifies that this inpainting step can be repeated in phases across patches of the image based on previous DDPM outputs).
Kim and Dey are considered analogous art, as they are both directed to machine learning models for image inpainting. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have integrated the teachings of Dey into Kim by using a diffusion/denoising process for the inpainting steps because the diffusion/denoising process produces high-quality synthetic image data (see Dey paragraphs 0039, 0051).
Regarding claim 6, Kim fails to disclose the limitations of claim 6.
Pertaining to the same field of endeavor, Dey discloses wherein generating the intermediate inpainting result comprises: identifying an intermediate diffusion timestep, wherein the first denoising phase includes a first plurality of denoising iterations performed prior to the intermediate diffusion timestep (see paragraphs 0039-0051 and 0058, where a DDPM / Denoising Diffusion Probabilistic Model is used to inpaint image regions via a denoising process, with a plurality of steps/timesteps "t"; paragraph 0085 specifies that this inpainting step can be repeated in phases across patches of the image based on previous DDPM outputs; thus the final timestep of any single inpainting step can be considered an intermediate diffusion timestep, as it comes after the plurality of denoising iterations of said inpainting step).
Kim and Dey are considered analogous art, as they are both directed to machine learning models for image inpainting. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have integrated the teachings of Dey into Kim by using a diffusion/denoising process for the inpainting steps because the diffusion/denoising process produces high-quality synthetic image data (see Dey paragraphs 0039, 0051).
Regarding claim 7, Kim fails to disclose the limitations of claim 7.
Pertaining to the same field of endeavor, Dey discloses the second denoising phase includes a second plurality of denoising iterations performed beginning from the intermediate diffusion timestep (the above final timestep of any single inpainting step cited in claim 6 would also precede all the denoising iterations of the next repeated inpainting step).
Kim and Dey are considered analogous art, as they are both directed to machine learning models for image inpainting. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have integrated the teachings of Dey into Kim by using a diffusion/denoising process for the inpainting steps because the diffusion/denoising process produces high-quality synthetic image data (see Dey paragraphs 0039, 0051).
Regarding claim 13, Kim in view of Dey discloses claim 13 as applied to claim 6 above.
Regarding claim 14, Kim in view of Dey discloses claim 14 as applied to claim 7 above.
Regarding claim 19, Kim fails to disclose the limitations of claim 19.
Pertaining to the same field of endeavor, Dey discloses identifying an intermediate diffusion timestep, wherein the intermediate inpainting result is generated using a first denoising process that includes a first plurality of denoising iterations performed prior to the intermediate diffusion timestep and the inpainting image is generated using a second denoising process that includes a second plurality of denoising iterations performed beginning from the intermediate diffusion timestep (see paragraphs 0039-0051, 0058, where a DDPM / Denoising Diffusion Probabilistic Model is used to inpaint image regions via a denoising process, with a plurality of steps/timesteps "t"; paragraph 0085 specifies that this inpainting step can be repeated in phases across patches of the image based on previous DDPM outputs; thus the final timestep of any single inpainting step can be considered an intermediate diffusion timestep, as it comes after the plurality of denoising iterations of said inpainting step, and precedes all the denoising iterations of the next repeated inpainting step).
Kim and Dey are considered analogous art, as they are both directed to machine learning models for image inpainting. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have integrated the teachings of Dey into Kim by using a diffusion/denoising process for the inpainting steps because the diffusion/denoising process produces high-quality synthetic image data (see Dey paragraphs 0039, 0051).
Claim 8 is rejected under 35 U.S.C. 103 as being unpatentable over Kim et al. (U.S. Publ. US-2023/0325985-A1) in view of Dey et al. (U.S. Publ. US-2026/0127730-A1), and further in view of Kudelski et al. (U.S. Publ. US-2022/0366544-A1).
Regarding claim 8, Kim in view of Dey fails to disclose the limitations of claim 8.
Pertaining to the same field of endeavor, Kudelski discloses the image generation model is trained to generate images at a first scale factor corresponding to the global context and at second scale factor corresponding to the local context (see paragraphs 0043-0055, where an inpainting preprocessing step includes determining an "image crop" / global context that is any subset of the image, and a "mask crop" / local context focused on a smaller region around the inpainting mask; each image and mask crop is assigned a size/scale factor and can also be downscaled by a specific factor).
Kim and Kudelski are considered analogous art, as they are both directed to machine learning models for image inpainting. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have integrated the teachings of Kudelski into Kim and Dey by using scale factors to generate the global/local contexts because doing so allows for determining the optimal crop size to balance context with processing load (see Kudelski paragraphs 0046, 0050-0051).
Claim 20 is rejected under 35 U.S.C. 103 as being unpatentable over Kim et al. (U.S. Publ. US-2023/0325985-A1) in view of Yuan et al. (U.S. Publ. US-2026/0094244-A1).
Regarding claim 20, Kim fails to disclose wherein the image generation model comprises a diffusion UNet. More specifically, in paragraphs 0034 and 0039, Kim discloses U-Nets and generative adversarial networks, but not diffusion U-Nets specifically.
Pertaining to the same field of endeavor, Yuan discloses wherein the image generation model comprises a diffusion UNet (see figure 5, stable diffusion u-net 510 and paragraphs 0032-0033, where the stable diffusion u-net can perform diffusion/denoising on noised input images; paragraphs 0017-0018 specify that the system can be used for inpainting).
Kim and Yuan are considered analogous art, as they are both directed to machine learning models for image inpainting. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have integrated the teachings of Yuan into Kim by using a diffusion u-net for both inpainting stages because doing so allows for completing occluded image content with high quality (see Yuan paragraph 0018).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to NICHOLAS JOHN HELCO whose telephone number is (703)756-5539. The examiner can normally be reached on Monday-Friday from 9:00 AM to 5:00 PM.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Matthew Bella, can be reached at telephone number 571-272-7778. 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 Patent Center. Status information for published applications may be obtained from Patent Center. Status information for unpublished applications is available through Patent Center for authorized users only. Should you have questions about access to Patent Center, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free).
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) Form at https://www.uspto.gov/patents/uspto-automated- interview-request-air-form.
/NICHOLAS JOHN HELCO/Examiner, Art Unit 2667 /MATTHEW C BELLA/Supervisory Patent Examiner, Art Unit 2667
Prosecution Insights