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 .
Response to Arguments
Applicant's arguments filed on 02/18/2026 have been fully considered but they are not persuasive.
Examiner notes that the amended claims appear to provide upscaling models for multiple resolutions but require only the target resolution for the end user device to be used. Examiner suggests elaborating on the claim limitations that link the two features and on a particular problem in the art that Applicant believes to solve by such limitations.
Applicant argues: “Applicant respectfully submits that Bampis does not suggest any use of super-resolution ML models, training super-resolution ML models, or generating a plurality of second reconstructed digital images that corresponds to the respective target resolution for the end user device, an as now recited above in amended claims 1, 9 and 15.”
Examiner notes that Wang teaches that super-resolution ML models are substantively identical to the models cited in Bampis. See reasons for rejection below.
Applicant argues: “Applicant respectfully submits that Wang does not suggest training each of a plurality of super-resolution machine learning (ML) models, nor generating a plurality of second reconstructed digital images that corresponds to the respective target resolution for the end user device, as recited in amended claims 1, 9 and 15.”
Examiner disagrees. The claim does not limit the target resolution for the end user device to any particular resolution. Bampis specifically teaches targeting the end user device as cited to Paragraphs 5 and 20, and Wang provides specific examples of end user device resolutions as cited in Paragraph 17. Examiner suggests elaborating on this limitation and its intended novelty.
Regarding arguments directed to the novelty of newly amended language, see treatment of the newly amended language in the updated reasons for rejection below.
35 USC § 101
Examiner has previously withdrawn the rejection of Claims 15-20 under 35 U.S.C. 101 in response to Applicant’s arguments submitted on 07/23/2025. Applicant has argued: “paragraph [0070] of Applicant's specification states "A computer readable storage medium, as used herein, is not to be construed as being transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission media (e. g. , light pulses passing through a fiber-optic cable), or electrical signals transmitted through a wire. "” This argument is persuasive, “computer readable storage medium” is a term that is defined in specification and excludes transitory embodiments from the definition.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with 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 claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1, 2-4, 6-20 recite a circular limitation: “generating a plurality of second reconstructed digital images with corrected image distortion by transforming the first reconstructed digital images using the plurality of trained super-resolution ML models of the ML decoder that upscales the plurality of second reconstructed digital images to a respective different second resolution that is greater than the first resolution and corresponds to the respective target resolution for the end user device;” In this case the plurality of second reconstructed digital images is an upscaled version of itself and thus lacks a proper definition in the claim. Examiner suggests amending the claim to recite “ML decoder that upscales the plurality of first reconstructed digital images.”
Claim Construction
Note that, for purposes of compact prosecution, multiple reasons for rejection may be provided for a claim or a part of the claim. The rejection reasons are cumulative, and Applicant should review all the stated reasons as guides to improving the claim language and advancing the prosecution toward an allowance.
Claim scope is not limited by claim language that suggests or makes optional but does not require steps to be performed by a method claim, or by claim language that does not limit an apparatus claim to a particular structure. However, examples of claim language, although not exhaustive, that may raise a question as to the limiting effect of the language in a claim are: (A) “adapted to” or “adapted for” clauses; (B) “wherein” clauses; and (C) “whereby” clauses. M.P.E.P. 2111.04. Other examples are where the claim passively indicates that a function is performed or a structure is used without requiring that the function or structure is a limitation on the claim itself. The clause may be given some weight to the extent it provides "meaning and purpose” to the claimed invention but not when “it simply expresses the intended result” of the invention. In Hoffer v. Microsoft Corp., 405 F.3d 1326, 1329, 74 USPQ2d 1481, 1483 (Fed. Cir. 2005). Further, during prosecution, claim language that may or may not be limiting should be considered non-limiting under the standard of the broadest reasonable interpretation. See M.P.E.P. 904.01(a); In re Morris, 127 F.3d 1048, 44 USPQ2d 1023 (Fed. Cir. 1997).
A claim containing a “recitation with respect to the manner in which a claimed apparatus is intended to be employed does not differentiate the claimed apparatus from a prior art apparatus” if the prior art apparatus teaches all the structural limitations of the claim. Ex parte Masham, 2 USPQ2d 1647 (Bd. Pat. App. & Inter. 1987). A recitation of the intended use of the claimed invention must result in a structural difference between the claimed invention and the prior art in order to patentably distinguish the claimed invention from the prior art. If the prior art structure is capable of performing the intended use, then it meets the claim.
Claim scope is not limited by claim language directed to a content of a signal but does not require steps to be performed, or by claim language that does not limit a claim to a particular structure, and thus does not require a separate reason for rejection. See, In re Lowry, 32 F.3d 1579, 1583-84, 32 USPQ2d 1031, 1035 (Fed. Cir. 1994); In re Ngai, 367 F.3d 1336, 1339, 70 USPQ2d 1862, 1864 (Fed. Cir. 2004); In re Gulack, 703 F.2d 1381, 1385, 217 USPQ 401, 404 (Fed. Cir. 1983). Where the claimed and prior art products are identical or substantially identical in structure or composition, or the claimed and prior art methods are identical or substantially identical processes, a prima facie case of either anticipation or obviousness has been established. In re Best, 562 F.2d 1252, 1255, 195 USPQ 430, 433 (CCPA 1977); M.P.E.P. 2112.01. “The functions corresponding to ‘processing,’ ‘receiving,’ and ‘storing’ are coextensive with a general purpose processor” In re Katz Interactive Call Processing Patent Litigation, 639 F.3d 1303, 1316, 97 USPQ2d 1737, 1747 (Fed. Cir. 2011). While substantive rejection of such language is provided below for purposes of compact prosecution, Examiner suggests rephrasing such claim language to recite limitations corresponding to the subject matter of the claim. Machine limitations should make clear that the use of the machine in the claimed process imposes a meaningful limitation on the claim’s scope. See MPEP 2106.01.
Note that “duplication of parts has no patentable significance unless a new and unexpected result is produced” In re Harza, 274 F.2d 669, 124 USPQ 378 (CCPA 1960).
"[E]ven though product-by-process claims are limited by and defined by the process, determination of patentability is based on the product itself. The patentability of a product does not depend on its method of production. If the product in the product-by-process claim is the same as or obvious from a product of the prior art, the claim is unpatentable even though the prior product was made by a different process." In re Thorpe, 777 F.2d 695, 698, 227 USPQ 964, 966 (Fed. Cir. 1985). See MPEP 2113(I).
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-4, 6-9, 10-11, 13-17, 19-20 are rejected under 35 U.S.C. 103 as being unpatentable over US 20230144735 to Bampis (“Bampis”) in view of US 20240273684 to Wang (“Wang”).
Regarding Claim 1: “A computer-implemented method comprising:
training each of a plurality of super-resolution machine learning (ML) models, based on training image data and a respective one of multiple target resolutions, to apply a super-resolution transformation to correct image distortion in reconstructed digital images at the respective target resolution for an end user device, wherein the respective target resolution of each of the plurality of super-resolution ML models is different; (“Although not shown, any number of instances of the training application 120 can be configured to generate any number of other jointly trained CNN pairs associated with different scale factors. … based on the resolution of the source video 106 and the resolutions specified in any number of encoding points (not shown).” Bampis, Paragraphs 56, 58, 89.
Prior art also teaches super-resolution as output resolution that is higher than the source resolution “the training application 120 generates the upscaling CNN 240 that includes a resolution increasing stack … a preliminary image having the same resolution as the source image … the output of the resolution increasing stack is a reconstructed image having a resolution that is higher than the resolution of the preliminary image,” and thus produces a super resolution that is higher than the source resolution. See Bampis, Paragraphs 81, 89 and Figure 2. See similarly in Wang, Paragraph 36 and the additional treatment of super-resolutions below.)
receiving, by a ML decoder that includes the plurality of trained super-resolution ML models, encoded image data, (“Upon receiving the chunk of the encoded video, the endpoint application decodes” Bampis, Paragraphs 19-20. “an endpoint application can identify, via metadata, the trained downscaling CNN used to generate an encoded video. The endpoint application can then identify and use the corresponding trained upscaling CNN to generate a corresponding reconstructed video that has an increased visual quality level” which indicates that the endpoint application includes a plurality of upscaling CNNs from which one needs to be identified. Bampis, Paragraph 12 and plurality of trained CNNs in Paragraphs 56, 58, 89.)
[encoded image data] that was downscaled from a first resolution of input digital images, wherein the encoded image data was generated by encoding first input digital images using a machine learning (ML) encoder; (First note that this element does not limit the claimed method to performing any particular steps. Describing the received data by the preferred process of generating it, does not limit the received data to be generated by only that process. See Claim Construction section above. Cumulatively, note that the ML nature of the encoders or decoders is not material to training or decoding super resolution in the present claim. See Specification, Paragraph 59.
Cumulatively note that prior art also performs this method in the context of operating on signals whose resolution was previously downscaled during encoding: “a first source image having a first resolution to generate a first downscaled image having a second resolution; … In practice, a typical encoding pipeline downscales the source video to multiple lower resolutions and then encodes the source video and each of the downscaled lower-resolution videos” Bampis, Paragraphs 11, 20. Here, encoding can based on machine learning: “a trained downscaling convolutional neural network (CNN) and a corresponding trained upscaling CNN can be implemented in combination with one another within a video encoding system” Bampis, Paragraph 12 and similarly in Wang, Paragraph 39.)
generating first reconstructed digital images by decoding the encoded image data using the ML decoder at the resolution corresponding to the ML encoder; and (“the endpoint application decodes the chunk and then optionally upscales the resulting decoded chunk to generate a corresponding chunk of reconstructed video having the same resolution as the client device display.” Bampis, Paragraphs 20, 12, and Figs. 2-3. As noted in Bampis Paragraph 12 and Wang Paragraph 39 above, the received image data is the type of data that can be produced by an ML encoder.)
generating a plurality of second reconstructed digital images with corrected image distortion by transforming the first reconstructed digital images using the plurality of trained super-resolution ML models of the ML decoder that upscales the plurality of second reconstructed digital images to a respective different second resolution that is greater than the first resolution and corresponds to the respective target resolution for the end user device; and (Under the broadest reasonable interpretation consistent with the specification and ordinary skill in the art, the plurality of reconstructed digital images can be an upscaling of a plurality of first reconstructed digital images (see treatment under Section 112 above), and the plurality of digital images can be frames of a video (See original Claim 8).
Prior art teaches this: “Upon receiving the chunk of the encoded video, the endpoint application decodes the chunk and then optionally upscales the resulting decoded chunk to generate a corresponding chunk of reconstructed video having the same resolution as the client device display. To affect the playback of the media title on the client device, the endpoint application plays back the different chunks of reconstructed video,” where a chunk of video corresponds to a frame of video which is a digital image, and the reconstructed video comprises a plurality of such reconstructed digital images. Bampis, Paragraphs 20, 66, and Fig. 1. As noted above, “a corresponding trained upscaling CNN can be implemented in combination with one another within a video encoding system.” Bampis, Paragraphs, 12, 60, and similarly in Wang, Paragraph 39.)
wherein the plurality of second reconstructed digital images has a higher image resolution compared with the first reconstructed digital images.” (For example, “the training application 120 generates the upscaling CNN 240 that includes a resolution increasing stack … The input of the preliminary layer stack is a downscaled image and the output of the preliminary layer stack is a preliminary image having the same resolution as the source image [first image] … the output of the resolution increasing stack is a reconstructed image having a resolution that is higher than the resolution of the preliminary image,” and thus produces a super resolution that is higher than the source resolution that corresponds to the first image. See Bampis, Paragraphs 81, 89 and Figure 2. See similarly in Wang, Paragraph 36 and the additional treatment of super-resolutions below.)
Bampis does not explicitly teach ”super-resolution machine learning (ML) models.” As noted above, Bampis teaches producing a reconstructed image with a resolution that is higher than the source image in Paragraphs 81 but does not explicitly describe this as super-resolution.
Wang explicitly describes this method as a super resolution upscaling, and it is typical in the context of using CNN to process encoded and decoded videos. See Wang, Paragraphs 15-17.
Therefore, before the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art to supplement the teachings of Bampis in a manner that produces “a second resolution greater than the first resolution” as taught in Bampis and/or Wang above, in order to reduce the resolution used in transmitting and processing the video signal. See Bampis Paragraph 19 and Wang Paragraph 17.
Finally, in reviewing the present application, there does not seem to be objective evidence that the claim limitations are particularly directed to: addressing a particular problem which was recognized but unsolved in the art, producing unexpected results at the level of the ordinary skill in the art, or any other objective indicators of non-obviousness.
Claim 9, “A system,” is rejected for reasons stated for Claim 1, and because prior art teaches: “a processor; and a memory having instructions stored thereon which, when executed on the processor, performs operations comprising:” (“Aspects of the present embodiments may be embodied as a system, method, or computer program product. … Furthermore, aspects of the present disclosure may take the form of a computer program product embodied in computer readable medium(s) having computer readable program code embodied thereon. … the computer readable storage medium would include the following: an electrical connection having wires, a portable computer diskette, a hard disk, a random access memory (RAM) …” Bampis, Paragraphs 160-161.)
Claim 15, “A computer program product,” is rejected for reasons stated for Claim 1, and because prior art teaches: “a computer-readable storage medium having computer-readable program code embodied therewith, the computer-readable program code executable by computer processors to perform operations, comprising:” (“Aspects of the present embodiments may be embodied as a system, method, or computer program product. … Furthermore, aspects of the present disclosure may take the form of a computer program product embodied in computer readable medium(s) having computer readable program code embodied thereon. … the computer readable storage medium would include the following: an electrical connection having wires, a portable computer diskette, a hard disk, a random access memory (RAM) …” Bampis, Paragraphs 160-161.)
Regarding Claim 2: “The computer-implemented method of claim 1, wherein
the ML encoder uses a first trained ML model to encode image and video data, (First note that the method of Claim 1 is a method of decoding, and does not include the encoder and the steps of encoding as a limitation on the method. For this reason, the method of this claim is rejected for reasons stated for Claim 1. Cumulatively, note that the ML nature of the encoders or decoders is not material to training super resolution of present claim. See Specification, Paragraph 59. Similarly, prior art indicates: “In embodiments, the encoder blocks and decoder blocks of FIG. 2 may use autoencoder structures (e.g., convolutional autoencoders).” Wang, Paragraph 39 and statement of motivation in Claim 1.)
the ML encoder encodes a second digital images generated using a media transformation layer to transform the first input digital images, and (“trained downscaling convolutional neural network (CNN) and a corresponding trained upscaling CNN can be implemented in combination with one another within a video encoding system to more effectively limit the diminution in visual quality of reconstructed videos when performing scaling operations.” Bampis, Paragraph 12. Note that “Each of the optional preliminary layer stack and the resolution decreasing stack includes learnable parameter values” Bampis, Paragraph 80.)
the second digital images are down-sampled from the first input digital images and have a lower resolution than the first digital images.” (“trained downscaling convolutional neural network (CNN) and a corresponding trained upscaling CNN can be implemented in combination with one another within a video encoding system to more effectively limit the diminution in visual quality of reconstructed videos when performing scaling operations.” Bampis, Paragraph 12.)
“the decoder uses a second trained ML model to decode image data previously encoded using the encoder.” (For example “a trained downscaling convolutional neural network (CNN) and a [second] corresponding trained upscaling CNN can be implemented” Bampis, Paragraph 12.
Cumulatively, Wang teaches the above claim feature in the context of encoding and decoding images with use of neural networks: “embodiments, the encoder blocks and decoder blocks of FIG. 2 may use autoencoder structures (e.g., convolutional autoencoders).” Wang, Paragraphs 39, 58. See statement of motivation in Claim 1.)
Regarding Claim 3: “The computer-implemented method of claim 2, wherein the media transformation layer uses a third trained ML model to transform the first input digital images and generate the second digital images.” (Note that this claim is directed to the function of an encoder which is not implemented by the claimed method of decoding and thus rejected for reasons stated for Claim 2. Cumulatively, prior art teaches: “trained downscaling [from first to second image] convolutional neural network (CNN) and a corresponding trained upscaling CNN can be implemented in combination with one another within a video encoding system to more effectively limit the diminution in visual quality of reconstructed videos when performing scaling operations.” Bampis, Paragraph 12. Note that “Each of the optional preliminary layer stack and the resolution decreasing stack includes learnable parameter values” Bampis, Paragraph 80.)
Regarding Claim 4: “The computer-implemented method of claim 2, wherein
the media transformation layer, the first trained ML model used by the ML encoder, and the second ML model used by the ML decoder are each respectively trained for reconstruction of images having a plurality of resolutions, and (See claim construction and rejection reasons in Claim 2. Cumulatively, prior art teaches: “any type of (previously) trained upscaler that implements any scale factor(s), or … trained downscaler that implements any scale factor(s) … includes multiple jointly trained CNN pairs associated with different scale factors,” where different scaling factors indicate different corresponding encoding and decoding resolutions. Bampis, Paragraphs 51 and 57. See treatment of ML encoders and decoders in Claim 1.)
each of a plurality of trained super-resolution ML models, is jointly trained with the ML encoder and ML decoder for reconstruction of images (See claim construction and rejection reasons in Claim 2. Cumulatively, prior art teaches: “a training application jointly trains a downscaling convolutional neural network (CNN) and an upscaling CNN … The training application then generates a training network that includes the downscaling CNN, the upscaling CNN, and an instance of the training up scaler that upscales images by the specified scale factor.” Bampis, Paragraphs 24-25. “includes multiple jointly trained CNN pairs associated with different scale factors,” where different scaling factors indicate different corresponding encoding and decoding resolutions. Bampis, Paragraph 57. This functionality can be integrated with a CNN encoder and decoder in a single neural network as noted in Wang, Paragraphs 15, 36. See statement of motivation in Claim 1.)
having the respective target resolution for the end user device.” (“the endpoint application executes one or more upscalers on the chunk of decoded video to generate the chunk of reconstructed video having the same resolution as the client device display” Bampis, Paragraphs 5, 19, 55, and Figure 1.)
Regarding Claim 6: “The computer-implemented method of claim 2, wherein
the first trained ML model used by the ML encoder is one of a plurality of trained encoder ML models, each of the plurality of trained encoder ML models trained for reconstruction of images having a respective target resolution, … the second trained ML model used by the ML decoder is one of a plurality of trained decoder ML models, each of the plurality of trained decoder ML models trained for reconstruction of images having a respective target resolution, and (See claim construction and rejection reasons in Claim 2. Also note that the described plurality of models (other than the first and second) are not used by the encoder and decoder of the claims, and are not used in implementing the steps of the method of Claim 6. Cumulatively, prior art teaches: “any type of (previously) trained upscaler that implements any scale factor(s), or … trained downscaler that implements any scale factor(s) … includes multiple jointly trained CNN pairs associated with different scale factors,” where different scaling factors indicate different corresponding encoding and decoding resolutions. Bampis, Paragraphs 51 and 57. This functionality can be integrated with a CNN encoder and decoder to include each trained model as noted in Wang, Paragraphs 15, 36. See statement of motivation in Claim 1.)
the super-resolution ML model is one of a plurality of trained super-resolution ML models, each super-resolution ML model trained for reconstruction of images (See claim construction and rejection reasons in Claim 2. Also note that the described plurality of models (other than the one ML model) are not used in implementing the steps of this method. Cumulatively, prior art teaches: “a training application jointly trains a downscaling convolutional neural network (CNN) and an upscaling CNN … The training application then generates a training network that includes the downscaling CNN, the upscaling CNN, and an instance of the training up scaler that upscales images by the specified scale factor.” Bampis, Paragraphs 24-25. “includes multiple jointly trained CNN pairs associated with different scale factors,” where different scaling factors indicate different corresponding encoding and decoding resolutions. Bampis, Paragraph 57.)
having the respective target resolution for the end user device.” (“the endpoint application executes one or more upscalers on the chunk of decoded video to generate the chunk of reconstructed video having the same resolution as the client device display” Bampis, Paragraphs 5, 19, 55, and Figure 1.)
Regarding Claim 7: “The computer-implemented method of claim 1, further comprising:
receiving the encoded image data at an electronic computing device using a communication network; and (“An encoded video associated with a given bitrate can be streamed to a client device without playback interruptions when the available network bandwidth is greater than or equal to that bitrate.” Bampis, Paragraph 19.)
presenting the second reconstructed digital images for display using a user-interface associated with the electronic computing device.” (“reconstructed video having the same resolution as the client device display. To affect the playback of the media title on the client device, the endpoint application plays back the different chunks of reconstructed video.” Bampis, Paragraph 19.)
Regarding Claim 8: “The computer-implemented method of claim 1, wherein, the first input digital images comprise frames in a digital video, and the second reconstructed digital images comprise frames in the digital video corresponding to the first digital images.” (“a trained downscaling convolutional neural network (CNN) and a corresponding trained upscaling CNN can be implemented in combination with one another within a video encoding system to more effectively limit the diminution in visual quality of reconstructed videos when performing scaling operations.” Bampis, Paragraph 12.)
Claim 10 is rejected for reasons stated for Claim 2 in view of the Claim 9 rejection.
Claim 11 is rejected for reasons stated for Claim 4 in view of the Claim 9 rejection.
Claim 13 is rejected for reasons stated for Claim 6 in view of the Claim 9 rejection.
Claim 14 is rejected for reasons stated for Claim 8 in view of the Claim 9 rejection.
Claim 16 is rejected for reasons stated for Claims2 in view of the Claim 15 rejection.
Claim 17 is rejected for reasons stated for Claim 4 in view of the Claim 15 rejection.
Claim 19 is rejected for reasons stated for Claim 6 in view of the Claim 15 rejection.
Claim 20 is rejected for reasons stated for Claim 8 in view of the Claim 15 rejection.
Claims 5, 12, 18 are rejected under 35 U.S.C. 103 as being unpatentable over US 20230144735 to Bampis (“Bampis”) in view of US 20240273684 to Wang (“Wang”) and in view of US 20210063200 to Kroepfl (“Kroepfl”).
Regarding Claim 5: “The computer-implemented method of claim 4, wherein a first super-resolution ML model of the plurality of super-resolution ML models is trained using transfer-learning, based on a previously trained second super-resolution ML model of the plurality of super-resolution ML models.” (See claim construction and rejection reasons in Claim 2.
Cumulatively, Bampis and Wang do not teach using transfer learning as the particular learning technique for its CNNs that implement the above function.
Kroepfl teaches the above claim feature in the context of using neural processing to perform data and image processing functions: “Training may be executed according to any classes of machine learning techniques, including, without limitation, classes such as: supervised training, semi-supervised training, unsupervised training, self-learning, reinforcement learning, federated learning, transfer learning, … and any variants or combinations therefor.” Kroepfl, Paragraph 261. See application to data of coded images and video in Paragraph 55.
Therefore, before the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art to supplement the teachings of Bampis and Wang to use transfer learning to implement the above claimed function as taught in Kroepfl, because it is a known and common method of training a CNN to perform data processing. See Kroepfl, Paragraph 261.
Finally, in reviewing the present application, there does not seem to be objective evidence that the claim limitations are particularly directed to: addressing a particular problem which was recognized but unsolved in the art, producing unexpected results at the level of the ordinary skill in the art, or any other objective indicators of non-obviousness.
Claim 12 is rejected for reasons stated for Claims 4 and 5 in view of the Claim 9 rejection.
Claim 18 is rejected for reasons stated for Claims 4-5 in view of the Claim 15 rejection.
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MIKHAIL ITSKOVICH whose telephone number is (571)270-7940. The examiner can normally be reached Mon. - Thu. 9am - 8pm.
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, Joseph Ustaris can be reached at (571)272-7383. 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.
/MIKHAIL ITSKOVICH/Primary Examiner, Art Unit 2483
Prosecution Insights