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 .
This is a 2nd action non-final rejection.
Response to Arguments
The rejection under 35 U.S.C. 112(b) has been withdrawn in light of the Applicant’s persuasive remarks on page 7 of the response filed on 6/2/2026.
The rejections under 35 U.S.C. 103 have been withdrawn in light of the Applicant’s persuasive remarks on pages 8-10 of the response filed on 6/2/2026. A new rejection has been proposed as described below.
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.
Claim(s) 1-4, 7, 10-14, 17, 20 are rejected under 35 U.S.C. 103 as being unpatentable over Kennett (US 20200389672 A1) in view of Ding (US 20250166126 A1), and Wimmer (US 20180365876 A1).
Regarding claim 1, Kennett disclose a method, comprising: receiving, by one or more processors, a compressed image comprising compression artifacts (Kennett, paragraph [0032], "As shown in FIG. 2, one or more video frames may include one or more decompression artifacts 214 as a result of compressing and decompressing the original video content 208").
While Kennett teaches training, by the one or more processors, an artificial intelligence (AI) model by fine-tuning a model (Kennett, paragraph [0022], “As used herein, a “machine learning model” refers to one or more computer algorithms or models (e.g., a classification model, a regression model) that can be tuned (e.g., trained) based on training input to approximate unknown functions”), they do not teach “fine-tuning a diffusion model”.
However, Ding teaches training, by the one or more processors, an artificial intelligence (AI) model by fine-tuning a diffusion model (Ding, paragraph [0005], "The technology described herein provides an improved training framework for a diffusion model used for a super resolution (SR) task.").
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to use diffusion models in place of Kennett’s models to perform video enhancements, as taught by Ding.
The suggestion/motivation for doing so would have been because diffusion models are more suitable for upscaling, as they specialize in sharpening image details rather than producing potential blurry results from traditional methods.
Further, one skilled in the art could have combined the elements as described above by known methods with no change in their respective functions, and the combination would have yielded nothing more than predictable results.
While Kennett in view of Ding teaches using training data comprising a plurality of training examples of compressed images with respective compression quality factors (Kennett, paragraph [0019], " For example, a compressed digital video may refer to a digital video (or series of video frames) that has been compressed using lossy or lossless compression algorithms. As a further example, in one or more embodiments, a compressed digital video is compressed using one or more block-oriented motion-compensation-based video compression standards. For instance, a compressed digital video may be compressed using formats including H264, H.265, MPEG-4, VP9, VP10, or any other encoding or compression format "), they do not teach “using training data comprising a plurality of training examples of compressed images annotated with respective compression quality factors”.
However, Wimmer teaches using training data comprising a plurality of training examples of compressed images annotated with respective compression quality factors (Wimmer, paragraph [0075], Fig. 3 below, "From the annotated ground truth, i.e. the annotated landmarks and structures in the training dataset 20, one or more of the following correspondent landmarks are extracted for model building, as illustrated by dataset 21 in FIG. 2 and FIG. 3 (see bright dots): two vertebral body center positions v.sub.j, center positions of middle d.sub.i, upper d.sub.i−1 and lower disc d.sub.i+1, and sampled points along the surface of the annotated cylinder", using the concept of labeling images with text).
PNG
media_image1.png
264
397
media_image1.png
Greyscale
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to annotate Kennett’s (in view of Ding) images with the compression format, as taught by Wimmer.
The suggestion/motivation for doing so would have been to gain a more thorough understanding of the video file for situations such as how it can be decoded.
Further, one skilled in the art could have combined the elements as described above by known methods with no change in their respective functions, and the combination would have yielded nothing more than predictable results.
Kennet in view of Ding and Wimmer discloses the diffusion model trained to perform super-resolution upscaling in accordance with an upscaling factor (Kennett, paragraph [0037], " For example, the super-resolution system 204 may include a super-resolution model (e.g., a machine learning model) trained to up-sample or otherwise increase the pixel resolution of one or more video frames"),
generating, by the one or more processors, an output image comprising fewer compression artifacts than the compressed image (Kennett, paragraph [0049], “ As shown in FIG. 3A, the denoising system 202 may apply a denoising model to the compressed video frame 302 based information from the segmentation mask 306 to generate a repaired video frame 310 in which one or more compression artifacts have been removed from the decompressed video frame 302”) and upscaled in accordance with the upscaling factor, the generating comprising providing the compressed image as input to the AI model and outputting, by the one or more processors, the output image on a display of one or more computing devices (Kennett, paragraph [0037], "Accordingly, the super-resolution system 204 may receive the repaired video frames 218 and generate high resolution video frames 220 to display via a display device 222").
Therefore, it would have been obvious to combine Kennett in view of Ding and Wimmer to obtain the invention as specified in claim 1.
Regarding claim 2, Kennett in view of Ding and Wimmer discloses the method of claim 1, wherein training the AI model by fine-tuning the diffusion model comprises: determining, by the one or more processors, a loss using an output of the diffusion model from the plurality of training examples and updating, by the one or more processors and in accordance with the loss, one or more model parameter values of the diffusion model (Ding, paragraph [0066], " The prediction is compared to the iteration specific diffusion data 230 or the adaption data 270 using a loss function.")*.
*As additionally evident by Wikipedia, neural networks have a loss function and
are trained to update their parameters in accordance to the loss functions.
PNG
media_image2.png
348
713
media_image2.png
Greyscale
Regarding claim 3, Kennett in view of Ding and Wimmer discloses the method of claim 2, wherein generating the output image comprises: adding noise, by the one or more processors, to the compressed image along a plurality of diffusion steps corresponding to diffusion operations to add noise to the compressed image and removing noise, by the one or more processors, from the noised compressed image along one or more denoising steps corresponding to denoising operations to remove noise and generate the output image (Ding, paragraph [0006], "Diffusion models work by corrupting the training data by progressively adding noise (e.g., Gaussian noise), slowly wiping out details in the training data until it becomes pure noise, and then training a neural network to reverse this corruption process.").
Regarding claim 4, Kennett in view of Ding and Wimmer discloses the method of claim 3, wherein the AI model is a pixel-space diffusion model (Ding, paragraph [0006], "Diffusion models work by corrupting the training data by progressively adding noise (e.g., Gaussian noise), slowly wiping out details in the training data until it becomes pure noise, and then training a neural network to reverse this corruption process.").
Regarding claim 7, Kennett in view of Ding and Wimmer discloses the method of claim 1, wherein receiving the training data comprises: receiving, by the one or more processors, an image compressed in accordance with a compression quality factor (Kennett, paragraph [0012], "The encoder system can generate a compressed digital video by applying a compression or encoding algorithm to the video content prior to transmitting the compressed video content to the client device"),
and generating, by the one or more processors, the compression quality factor as a label for the image (Wimmer, paragraph [0075], "From the annotated ground truth, i.e. the annotated landmarks and structures in the training dataset 20, one or more of the following correspondent landmarks are extracted for model building, as illustrated by dataset 21 in FIG. 2 and FIG. 3 (see bright dots): two vertebral body center positions v.sub.j, center positions of middle d.sub.i, upper d.sub.i−1 and lower disc d.sub.i+1, and sampled points along the surface of the annotated cylinder", as cited in claim 1, the images are annotated with the compression formats).
Regarding claim 10, Kennett in view of Ding and Wimmer discloses the method of claim 1, wherein each compressed image in the training data is lossily compressed (Kennett, paragraph [0019], " For example, a compressed digital video may refer to a digital video (or series of video frames) that has been compressed using lossy or lossless compression algorithms").
Claims 11-14, 17 corresponds to claims 1-4, 7, additionally reciting a system with one or more processors (Kennett, paragraph [0083], “The computer system 700 includes a processor 701”). Thus, they are rejected for the same reasons of obviousness as claims 1-4, 7.
Claim 20 corresponds to claim 1, additionally reciting one or more non-transitory
computer-readable storage media, storing instructions that when executed by one or more
processors (Kennett, paragraph [0089], “If implemented in software, the techniques may be realized at least in part by a non-transitory processor-readable storage medium comprising instructions that, when executed by at least one processor, perform one or more of the methods described herein”). Thus, it is rejected for the same reasons of obviousness as claim 1.
Claim(s) 5 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Kennett (US 20200389672 A1) in view of Ding (US 20250166126 A1), Wimmer (US 20180365876 A1), and in further view of Pan (US 20250014233 A1).
Regarding claim 5, Kennett in view of Ding and Wimmer discloses the method of claim 4.
Kennett in view of Ding and Wimmer does not teach “wherein removing noise from the noised compressed image along the one or more denoising steps comprises processing, by the one or more processors, the noised compressed image through a consistency model trained to generate the output image by evaluating a probabilistic flow ordinary differential equation (ODE)”.
However, Pan teaches wherein removing noise from the noised compressed image along the one or more denoising steps comprises processing, by the one or more processors, the noised compressed image through a consistency model trained to generate the output image by evaluating a probabilistic flow ordinary differential equation (ODE) (Pan, paragraph [0053], "The de-noising module 250 can execute a de-noising process by solving a deterministic probability-flow ordinary differential equation (ODE) instead of by a stochastic de-noising process (e.g., as represented by p.sub.θ(x.sub.k-1|x.sub.k) in FIG. 3)").
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to implement a model
that de-noises Kennett’s (in view of Ding and Wimmer) image through a PF-ODE, as taught by Pan.
The suggestion/motivation for doing so would have been to achieve de-noising in
fewer steps.
Further, one skilled in the art could have combined the elements as described above by known methods with no change in their respective functions, and the combination would have yielded nothing more than predictable results.
Therefore, it would have been obvious to combine Kennett in view of Ding, Wimmer, and in further view of Pan to obtain the invention as specified in claim 5.
Claim 15 corresponds to claim 5, additionally reciting a system (Kennett, paragraph [0083], “The computer system 700 includes a processor 701”). Thus, it is rejected for the same reasons of obviousness as claim 5.
Allowable Subject Matter
Claims 6, 8, 16, 18 objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to WAYNE ZHANG whose telephone number is (571) 272-0245. The examiner can normally be reached Monday-Friday 10:00-6:00 EST.
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, Ms. Sumati Lefkowitz can be reached on (571) 272-3638. 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.
/WAYNE ZHANG/Examiner, Art Unit 2672
/SUMATI LEFKOWITZ/Supervisory Patent Examiner, Art Unit 2672
Prosecution Insights