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 .
Information Disclosure Statement
The information disclosure statement (IDS) submitted on November is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is considered by examiner.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The 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 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Mao et al (Generation of Defective Lithium Battery Electrode Samples Based on Diffusion Models) in view of Liu et al (Assigned MURA Defect Generation Based on Diffusion Model).
Regarding Claim 1, Mao et al teach a method (method of using a Denoising Diffusion Probabilistic Model (DDPM); III. Method Diffusion model) comprising:
generating, by a processor via a diffusion model, a noisy image from a defect-free image (the diffusion model is used to iteratively add noise to a real normal input image (defect-free image of a image set) in the Forward Process; Fig 2, 6 and III.B. Diffusion Model, IV.A Experiment Datasets, IV.B Results ¶ 3);
generating, by the processor via the diffusion model, a sampled defect image and a sampled defect-free image from the noisy image (the reverse diffusion process is used to iteratively remove noise of the defect sample image and normal sample image based on the associated noise image from the forward process to restore the image; Fig 2, 5, 6 and III.B. Diffusion Model, IV.A Experiment Datasets, IV.B Results ¶ 3);
Mao et al does not explicitly disclose generating, by the processor, a mask based on the sampled defect image and the sampled defect-free image; generating, by the processor, a synthetic defect image by generating an additional sampled defect image based on the noisy image and the mask; and transmitting, by the processor, the synthetic defect image.
Liu et al is analogous art pertinent to the technological problem addressed in the current application and teaches generating, by the processor, a mask based on the sampled defect image and the sampled defect-free image (a mask can be generated based on a defect region (the defect is determined by a comparison to a defect free (random) region of the image) and an mask is generated for the random region as well for the model to distinguish between defect and non-defect; Fig 3 and 3.3 Defect Generation based on Conditional Diffusion ¶ 1-2);
generating, by the processor, a synthetic defect image by generating an additional sampled defect image based on the noisy image and the mask (the mask region is used with the stored feature images to generate additional (synthetic “fake”) defect images; Fig 4 and 3.4 Feature Reservation ¶ 2); and
transmitting, by the processor, the synthetic defect image (the synthetic “fake” images are stored and used for additional training of the diffusion model; Fig 4 and 3.4 Feature Reservation ¶ 2, 5.1 Conclusions).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the current application to combine the teachings of Mao et al with Liu including generating, by the processor, a mask based on the sampled defect image and the sampled defect-free image; generating, by the processor, a synthetic defect image by generating an additional sampled defect image based on the noisy image and the mask; and transmitting, by the processor, the synthetic defect image. By using a mask associated with a defect, a specific area of the image may be changed and used for additional training data, thereby improving the diffusion model efficiently with reduced computational requirements, as recognized by Liu et al (3.3 Defect Generation based on Conditional Diffusion ¶ 1).
Regarding Claim 2, Mao et al in view of Liu et al teach the method of claim 1 (as described above), wherein the defect-free image comprises a real defect-free image of a target product (Mao et al, the image datasets include real images of a correctly folded lithium battery electrode; Fig 6 and IV.A. Datasets).
Regarding Claim 3, Mao et al in view of Liu et al teach the method of claim 1 (as described above), wherein the generating the sampled defect image comprises sampling the diffusion model based on a first class label set to a defect label (Mao et al, the diffusion model includes generation of (diffusion) defect images based on real defective electrode folding images with the defects based on classification (labeling to identify normal from defect); Fig 3-5 and IV.B. Experiments Results ¶ 3).
Regarding Claim 4, Mao et al in view of Liu et al teach the method of claim 3 (as described above), wherein the generating the sampled defect-free image comprises sampling the diffusion model based on a second class label set to a defect-free label (Mao et al, the diffusion model includes generation of (diffusion) normal images based on original real normal electrode images with the normal images identified based on classification (labeling to identify normal from defect); Fig 3-5 and IV.B. Experiments Results ¶ 3).
Regarding Claim 5, Mao et al in view of Liu et al teach the method of claim 4 (as described above), wherein the generating the sampled defect image and the sampled defect-free image comprises sampling the diffusion model for a set of time steps in an iterative process (Liu et al, the diffusion process for forward (and therefore backward) passes are performed over a series of time steps t; 3.1 Image Generation with DDPM ¶ 1-2).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the current application to combine the teachings of Mao et al with Liu including the denoising is performed for the additional sampled defect image for a set of time steps the generating the sampled defect image and the sampled defect-free image comprises sampling the diffusion model for a set of time steps in an iterative process. By using time steps the data is iteratively diffused of random noise, thereby improving the diffusion model efficiently with reduced computational requirements by continuous step improvements, as recognized by Liu et al (3.3 Defect Generation based on Conditional Diffusion ¶ 1).
Regarding Claim 6, Mao et al in view of Liu et al teach the method of claim 1 (as described above), wherein the generating the mask comprises determining a difference between the sampled defect image and the sampled defect-free image (Liu et al a mask can be generated based on a defect region (the defect is interpreted to be determined by a comparison to a defect free region of the image) and an mask is generated for the random region as well for the model to distinguish between defect and non-defect; Fig 4 and 3.3 Defect Generation based on Conditional Diffusion ¶ 1-2).
Regarding Claim 7, Mao et al in view of Liu et al teach the method of claim 1 (as described above), wherein the generating the synthetic defect image comprises generating a defect within a location defined by the mask and overlaying the mask over the additional sampled defect image (Liu et al, the masks are used to generate synthetic defect images with the defect only in specific mask positional locations; Fig 4 and 3.4 Feature Reservation ¶ 2, 5.1 Conclusions).
Regarding Claim 8, Mao et al in view of Liu et al teach the method of claim 1 (as described above), wherein the generating the synthetic defect image comprises replacing a background of the synthetic defect image outside of a location defined by the mask (Liu et al, the background of the synthetic images include regions identified as regions to generate normal background (shown as black boxes); Fig 4(c) and 3.4 Feature Reservation ¶ 2, 5.1 Conclusions).
Regarding Claim 9, Mao et al in view of Liu et al teach the method of claim 8 (as described above), wherein the replacing the background of the synthetic defect image is performed based on the defect-free image (Liu et al, the background of the synthetic images include regions identified as regions to generate normal background (shown as black boxes) based on the normal real image data; Fig 4(c) and 3.4 Feature Reservation ¶ 2, 5.1 Conclusions).
Regarding Claim 10, Mao et al in view of Liu et al teach the method of claim 1 (as described above), wherein the synthetic defect image comprises a synthetic defect image of a target product and the defect-free image of the target product (Mao et al, training data includes normal images and augmented defective images of the electrodes; IV.B. Experiments Results ¶ 3).
Regarding Claim 11, Mao et al in view of Liu et al teach the method of claim 1 (as described above), wherein the generating the synthetic defect image comprises denoising an amount of noise within a location defined by the mask (Liu et al, the masks are used to generate synthetic defect images with the defect only in specific mask positional locations; Fig 4 and 3.4 Feature Reservation ¶ 2, 5.1 Conclusions).
Regarding Claim 12, Mao et al in view of Liu et al teach the method of claim 11 (as described above), wherein the denoising is performed for the additional sampled defect image for a set of time steps (Liu et al, the diffusion process for forward (and therefore backward) passes are performed over a series of time steps t; 3.1 Image Generation with DDPM ¶ 1-2).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the current application to combine the teachings of Mao et al with Liu including the denoising is performed for the additional sampled defect image for a set of time steps. By using time steps the data is iteratively diffused of random noise, thereby improving the diffusion model efficiently with reduced computational requirements by continuous step improvements, as recognized by Liu et al (3.3 Defect Generation based on Conditional Diffusion ¶ 1).
Regarding Claim 13, Mao et al in view of Liu et al teach the method of claim 1 (as described above), wherein the diffusion model is trained based on a real defect-free image and a real defect image (Mao et al, initial training and testing is performed using (real) normal images and (real) defective images; Fig 5, 6 and IV.B. Experiments Results ¶ 3).
Regarding Claim 14, Mao et al in view of Liu et al teach the method of claim 13 (as described above), wherein the real defect-free image and the real defect image are associated with a source product (Mao et al, the image datasets are images of a folded defect in a lithium battery electrode; Fig 5, 6 and IV.A. Datasets).
Regarding Claim 15, Mao et al teach a device (computer operating system; IV. Experiments) comprising: one or more processors (GPU RTX3090Ti to execute instructions using Python version 3.6 to execute diffusion model; IV. Experiments) that are configured to perform: steps identical to claim 1 (as discussed above).
Regarding Claim 16, Mao et al in view of Liu et al teach the device of claim 15 (as described above), wherein further steps identical to claim 3 (as discussed above).
Regarding Claim 17, Mao et al in view of Liu et al teach the device of claim 16 (as described above), wherein further steps identical to claim 4 (as discussed above).
Regarding Claim 18, Mao et al in view of Liu et al teach the device of claim 17 (as described above), wherein further steps identical to claim 5 (as discussed above).
Regarding Claim 19, Mao et al in view of Liu et al teach the device of claim 15 (as described above), wherein further steps identical to claim 6 and 7 (as discussed above).
Regarding Claim 20, Mao et al teach a system (computer operating system; IV. Experiments) comprising: a processing circuit (GPU RTX3090Ti contains a processor; IV. Experiments); and a memory storing instructions, which, based on being executed by the processing circuit (GPU RTX3090Ti contains a (GDDR6X) memory to store diffusion model instructions, executed on processor; IV. Experiments), cause the processing circuit to perform: steps identical to claim 1 (as discussed above).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Ridder et al (SEMI-DiffusionInst: A Diffusion Model Based Approach for Semiconductor Defect Classification and Segmentation) teach a system and method for defect inspection of semiconductors based on defect pattern recognition using a diffusion model and includes use of a bounding box and segmentation techniques for identifying the region of interest.
Clever et al (US 2025/0061583) teach a method and system for synthetic image generation and augmentation using a diffusion model.
NVIDISA GeForce RTX 3090Ti provides key specifications to serve as evidence the graphic card contains memory and processor components.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KATHLEEN M BROUGHTON whose telephone number is (571)270-7380. The examiner can normally be reached Monday-Friday 8:00-5:00.
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, John Villecco can be reached at (571) 272-7319. 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.
/KATHLEEN M BROUGHTON/Primary Examiner, Art Unit 2661
Prosecution Insights