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 .
DETAILED ACTION
This Office Action is in response to the application 19/118,933 filed on 04/07/2025.
Claims 1 – 10, 17 and 19 - 26 have been examined and are pending in this application.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 04/07/2025. The submission is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Specification
The lengthy specification has not been checked to the extent necessary to determine the presence of all possible minor errors. Applicant’s cooperation is requested in correcting any errors of which applicant may become aware in the specification.
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, 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.
Claim 17 is 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.
Claim 17 recites “17. (Currently amended) A non-transitory computer readable medium storing a bitstream comprising scalable neural-network based coded data representative of image data samples for a neural network-based vision inference processing and associated metadata, wherein the associated metadata comprises at least one of:”. Claim 17 is directed to a non-transitory computer-readable medium for storing a bitstream clauses that appear to describe how the bitstream is processing or generating. These elements or steps are not performed by an intended computer, and the bitstream is not a form of programming that causes functions to be performed by an intended computer. This shows that the computer-readable medium merely serves as support for storing the bitstream and provides no functional relationship between the steps/elements that describe the generation of the bitstream and intended computer system. Therefore, those claim elements are not given patentable weight. Patentable weight is given to data stored on a computer-readable medium when there exists a functional relationship between the data and its associated substrate. See MPEP 2111.05 III. For example, if a claim is drawn to a computer-readable medium containing programming/ instructions, a functional relationship exists if the programming “performs some function with respect to the computer with which it is associated.” However, if the claim recites that the computer-readable medium merely serves as a storage for information or data that is not meant for being executed, no functional relationship exists and the information or data is not given patentable weight.
The Examiner suggests that the claim be amended so that it is directed to a functional relationship. For example, in this particular case, the claim should instead be recited as “A non-transitory computer-readable storage medium for storing a computer program and a bitstream, wherein when executed by a processor, the computer program causes the processor to process the bitstream and associated metadata, the processor configured to.. or correct as needed.
Claim Rejections - 35 USC § 102
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 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 and 17 are rejected under 35 U.S.C. 102(a)(1) as being by Hyomin et al. (“Scalable Image Coding for Humans and Machines” dated Jan 13, 2022).
Regarding claim 1, Hyomin discloses: “a method, comprising:
obtaining a tensor of reconstructed data representative of image data samples partially reconstructed from a base layer of a scalable bitstream, the tensor of reconstructed data comprises a first number of channels of two-dimensional data, reconstructed data having an input spatial resolution [see Fig. 1; page: 2; section: II. PRELIMINARIES AND RELATED WORK; DNN-based image compression approaches [14]–[18] map the input image X into a latent representation Y ∈ RN×M×C, as shown in Fig. 2(a), where C is the number of latent feature channels and N ×M is the resolution of each channel. The latent representation Y is compressed, and later decoded to produce the reconstructed input X ≈ X. Several multi task DNNs [29]–[31] are based on a similar approach, but in addition to input reconstruction X, they enable computer vision inference T (e.g., classification, object detection, etc.) from the latent space without input reconstruction, as shown in Fig. 2(b). And section: III; In developing a scalable DNN-based multi-task compression system, the first step is to choose the backbone that can support the tasks of interest. Since one of the tasks is input reconstruction, we select a recent backbone from [16], which has shown competitive performance against conventional image codecs. Its operation can be described as [14]–[18]: Y =ga(X;φ) Y =Q(Y) X=gs(Y;θ) (1) where ga and gs are the analysis and synthesis transforms, respectively, φ and θ are their parameters, Q represents quantization, and the lossless operations of entropy coding and decoding have been omitted from (1). Let T be a machine vision inference task that needs to be derived from the input image X. Since X ≈ X at high enough rate, this inference can also be derived from X.1 Let V be a DNN that computes T, which operates as: F =V(front-end)(X;ψ) T =V(back-end)(F;ρ) (2) where ψ and ρ are the parameters of the DNN’s front-end and back-end, respectively. V (front-end) consists of the DNN’s input layer and a number of subsequent layers, and generates the intermediate feature tensor F. V (back-end) consists of the remaining DNN layers and produces the inference output T];
applying to the tensor of reconstructed data a neural network-based feature synthesis processing to generate a tensor of input feature representative of a feature of image data samples, the tensor of input feature comprises a second number of channels of two- dimensional data [see page: 4; section: C; The base layer Y1 will support the base computer vision task, say object detection, for example. However, Y1, in gen eral, does not match any of intermediate latent spaces of any of the object detectors. Hence, we introduce the latent space transform (LST) to map Y1 to an intermediate latent space of an object detector we want to use at the decoder. Specifically, let F(l) 1 be the feature tensor at the output of the l-th layer of a chosen object detector V1, that is, F(l) 1 =V(front-end) ( X,ψ), as in (2). Then the goal of the base LST is to map Y1 to 1 F(l) 1 . More generally, for the j-th task, we use { Y1,..., Yj} from the encoded latent space, and the goal of j-th LST is to map {Y1,..., Yj} to F(l) j =V(front-end) j the DNN implementing the j-th task];
resizing the tensor of input feature to generate a tensor of output feature, the tensor of output feature comprises a third number of channels of two-dimensional data, wherein output feature data have an output spatial resolution that is an arbitrary ratio of the input spatial resolution [see fig. 3, fig. 5; section II: "DNN-based image compression approaches [14]-[16] - map the input image X into a latent representation Y € RNXMXC, as shown in Fig. 2 (a), where C is the number of latent feature channels and NxM is the resolution of each channel. The latent representation Y is compressed, and later decoded to produce the reconstructed input X^ ≈ X. Several multitask DNNs [26]-[28] are based on a similar approach, but in addition to input reconstruction X^, they enable computer vision inference T (e.g., classification, object detection, etc.) from the latent space In particular, it will be seen that the information needed for T is a subset of that needed for X^, which is the main motivation behind our proposed latent-space scalability Only Y1 needs to be decoded to perform inference task T, whereas both {Y1,Y2} are used to reconstruct X^", section III.B:" The latent representation Y is split into sub-latents {Y1,Y2}, where Y1 = {Y1,Y2, Yi} represents the base layer and Y2 = {Yi+1,Yi+2, YC} represents the enhancement layer." section III.C: A RB with upsampling factor rk (shown as 1 rk) increases spatial dimension by rk and also applies the inverse GDN [40]. Upsampling factors rk are chosen based on the dimensions of { Y1,..., Yj} and F(l)]; and
applying to the tensor of output feature, a neural network-based vision inference processing to generate a collection of inference results [see Fig. 3; section: III];
wherein resizing the tensor of input feature comprises applying at least one interpolation filter to the tensor of input feature to adapt at least a dimension of the tensor of input feature to the neural network-based vision inference processing [see (fig. 3, 5, section IV. B: Two-layer network: Our two layer network supports object detection in the base layer using theYOLOv3 [27] back end. We first evaluate the object detection performance on the COCO2014validation set [57], which includes about 5,000 images. Since most vision networks resize the input to a specific resolution before processing, we also resize input images to 512×512 using bilinear interpolation without letterboxing, in order to generate (viaLST) a feature tensor F(13) 1 ∈R64×64×256thatcanbedirectlyfedintotheYOLOv3 back-end)].
Regarding claim 2 and 17, claim 2 and 17 is rejected under the same art and evidentiary limitations as determined for the method of claim 1.
Allowable Subject Matter
Claims 3 – 10 and 19 - 26 are 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
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
WO 2022/128137 A1.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Masum Billah whose telephone number is (571)270-0701. The examiner can normally be reached Mon - Friday 9 - 5 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, Jamie J. Atala can be reached at (571) 272-7384. 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.
/MASUM BILLAH/Primary Patent Examiner, Art Unit 2486
Prosecution Insights