DETAILED ACTION
1. The communication is in response to the application received 07/07/2025, wherein claims 1-17 are pending are examined as follows.
Notice of Pre-AIA or AIA Status
2. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Priority
3. Receipt is acknowledged of certified copies of papers required by 37 CFR 1.55.
Information Disclosure Statement
4. The information disclosure statements (IDS) were submitted on 07/07/2025. The submissions are in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statements are being considered by the examiner.
Claim Objections
5. Claim 6 is objected to because of the following informalities: claim 6 recites “The method of claim 1, wherein the first portion and second portion do not include a hierarchical representation of an input to the first portion.” (emphasis added), however, claim 1 does not mention a “first portion” or a “second portion”. Since these appear in claim 5, it is believed claim 6 should depend on claim 5. Please check and update accordingly. Appropriate correction is required.
Claim Rejections - 35 USC § 103
6. 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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1-2, 7-8, and 12-17 are rejected under 35 U.S.C. 103(a) as being unpatentable over Kim et al. US 2023/0105112 A1, in view of Rao et al. US 2022/0253220 A1, hereinafter referred to as Kim and Rao, respectively.
Regarding claim 1, Given the broadest reasonable interpretation (BRI), Kim teaches and/or suggests “A method of encoding a tensor for a single frame into a bitstream [See for e.g. the encoding process of fig. 3 and associated text (e.g.¶ 0079) with respect to encoding feature maps], the method comprising: producing coefficients for the tensor using a plurality of basis vectors, [See for e.g. ¶0107 regarding the use of PCA from which optimum transform vectors can be derived for a transform unit group (i.e. channels of a feature map). Optimum transform vectors can be construed as basis vectors since these are obtained via PCA. Please note, basis vector sets are also discussed throughout Kim (e.g. ¶0121-¶0122, fig. 19). ¶0113 further shows transform coefficients can be outputted for the feature maps] the coefficients forming one feature map for each basis vector of the plurality of basis vectors [Fig. 9, for e.g., shows the outputted P2, P3, P4, and P5 transform coefficients corresponding to their respective feature maps based on the derived optimum transform vector(s)]; producing an assignment that groups coefficients according to the respective one of the basis vector for the coefficients [Since an assignment groups coefficients, and given the BRI of “assignment”, see Kim’s feature map information packer 150 which includes grouping unit 151 and arrangement unit 152 (e.g. fig. 13). Said grouping unit can group transform coefficients into one or more data groups (e.g. ¶0121). Once grouped, the data may then be arranged as noted in for e.g. in ¶0122]; encoding the coefficients into a feature frame bitstream such that each group of coefficients is decoded independently [Figs. 1 and 9 illustrate encoding a feature frame bitstream. As to being decoded independently, see fig. 15 which shows groups of coefficients P2, P3, P4, and P5 that can be decoded independently of each other to yield corresponding reconstructed feature maps]; and encoding the assignment into the bitstream [Additional information may be transmitted in the bitstream. For e.g., said information may describe whether arrangement is performed on data that is derived from a feature map/tensor (e.g. ¶0123-¶0124)] and setting a subpicture arrangement based on the assignment to encode the tensor.” [Kim however does not appear to address a sub-picture arrangement as claimed. Please see Rao below for support] Although Kim’s feature map information packer 150 (e.g. ¶0121-¶0122) is deemed relevant regarding “the assignment to encode the tensor” (see above), Kim does not appear to address “setting a subpicture arrangement” based on said assignment. Rao on the other hand from the same or similar field of endeavor is relied on to teach and/or suggest “and setting a subpicture arrangement based on the assignment to encode the tensor.” (emphasis added) [Given the BRI of the limitation, Rao’s uncompressed feature maps are grouped into tiles (here construed as subpictures) which in turn can be compressed on a tile-by-tile basis based on a predetermined format of subdivision (e.g. ¶0061-¶0063)] Given Rao’s teachings above, it would have therefore been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify teachings of Kim for encoding/decoding feature maps (e.g. abstract), to add the neural processing techniques of Rao as above for arranging compressed feature map data in unitary storage spaces to help reduce the number of memory accesses and to allow parts of the feature map to be retrieved and processed flexibly (e.g. ¶0059).
Regarding claim 2, Kim and Rao teach and/or suggest all the limitations of claim 1, and are analyzed as previously discussed with respect to that claim. Although Kim discusses grouping and arranging transform coefficients via feature map information packer 150 (fig. 13), which can be construed as an “assignment” given its BRI, Kim does not appear to address all of the features of claim 2. Rao, on the other hand, from the same or similar field of endeavor, teaches and/or suggests “wherein when it is determined to update the assignment, a new subpicture structure is commenced.” [The phrase “update the assignment” is taken to mean rearranging the order in which a feature map is processed. As such, please refer to ¶0023 of Rao, which describes a flexible processing order of an initial arrangement of feature map data. By re-arranging the order for processing the tiles in a feature map, this suggests a new tile structure can be initiated given the BRI of “a new subpicture structure is commenced”] The motivation for combining Kim and Rao has been discussed in connection with claim 1, above.
Regarding claim 7, Kim and Rao teach and/or suggest all the limitations of claim 1, and are analyzed as previously discussed with respect to that claim. Kim further teaches/suggests “wherein the assignment is produced using coefficients obtained from the feature frame.” [Since an assignment groups coefficients, and given the BRI of “assignment”, see Kim’s feature map information packer 150 which includes grouping unit 151 and arrangement unit 152 (e.g. fig. 13). Said grouping unit can group transform coefficients into one or more data groups (e.g. ¶0121). Once grouped, the data may then be arranged as noted in for e.g. in ¶0122]
Regarding claim 8, claim 8 is rejected under the same art and evidentiary limitations as determined for the method of Claim 1 since encoding and decoding are inverse operations of each other that allows a decoder to decompress an encoded tensor/feature mask at a receiving device. Regarding encoding and decoding, please see figs. 3 and 14 of Kim.
Regarding claim 12, claim 12 is rejected under the same art and evidentiary limitations as determined for the method of Claim 1.
Regarding claim 13, claim 13 is rejected under the same art and evidentiary limitations as determined for the method of Claim 1. As to the claimed hardware and software, please see fig 20 of Kim for support.
Regarding claim 14, claim 14 is rejected under the same art and evidentiary limitations as determined for the method of Claim 1. As to the claimed hardware and software, please see fig 20 of Kim for support.
Regarding claim 15, claim 15 is rejected under the same art and evidentiary limitations as determined for the method of Claim 1 since encoding and decoding are inverse operations of each other that allows a decoder to decompress an encoded tensor/feature mask at a receiving device. Regarding encoding and decoding, please see figs. 3 and 14 of Kim.
Regarding claim 16, claim 16 is rejected under the same art and evidentiary limitations as determined for the method of Claim 1 since encoding and decoding are inverse operations of each other that allows a decoder to decompress an encoded tensor/feature mask at a receiving device. As to the claimed hardware and software, please see fig 20 of Kim for support.
Regarding claim 17, claim 17 is rejected under the same art and evidentiary limitations as determined for the method of Claim 1 since encoding and decoding are inverse operations of each other that allows a decoder to decompress an encoded tensor/feature mask at a receiving device. As to the claimed hardware and software, please see fig 20 of Kim for support.
Claim 4 is rejected under 35 U.S.C. 103(a) as being unpatentable over Kim, in view of Rao, and in further view of Han et al.[VCM track 1] EE1.1 multi-scale feature compression anchor generation for object detection on OpenImage V6 dataset, International Organization for Standardization, ISO/EC JTC 1/SC 29/WG 2 MPEG Technical Requirements, ISO/IEC JTC 1/SC 29/WG 2 m59576, April 2022 (see IDS dated 07/07/2025), hereinafter referred to as Han.
Regarding claim 4, Kim and Rao teach and/or suggest all the limitations of claim 1, and are analyzed as previously discussed with respect to that claim. However, Kim and Rao do not appear to address the features of claim 4. Han, on the other hand, from the same or similar field of endeavor, teaches and/or suggests “wherein the tensor is the result of fusing a plurality of tensors forming a hierarchical representation together.” [Given ¶00092 of the filed specification shows one type of multi-scale feature fusion module may serve as a tensor combiner, please refer to pg. 1 Sect. 1 “Anchor generation”, which shows that multi-scale feature compression (MSFC) consists of a multi-scale feature fusion module (MSFF)] Given Han’s teachings above, it would have therefore been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify teachings of Kim and Rao, to add the
techniques of Han regarding enhanced MSFR based on multi-scale feature compression (MSFC) as above such that the data size of the feature map can be greatly reduced with little loss of accuracy (abstract).
Claims 5-6 are rejected under 35 U.S.C. 103(a) as being unpatentable over Kim, in view of Rao, and in further view of Wu et al. US 2024/0206907 (with reference to WO 2022/232701 A1), hereinafter referred to as Wu.
Regarding claim 5, Kim and Rao teach and/or suggest all the limitations of claim 1, and are analyzed as previously discussed with respect to that claim. Kim further teaches/suggests “wherein the tensor is the sole dividing point of a network that is split into a first portion and second portion.” [See Kim’s fig. 4 and ¶0083-¶0084, where the feature map extraction unit may have a feature pyramid network structure or a structure with multiple convolutional layers combined. Tensors are depicted by feature maps between layers which can be construed as dividing points of the network] Although Kim (not Rao) reasonably addresses the aforementioned features given their BRI, the work of Wu from the same or similar field of endeavor is brought in to provide further support that aligns with the filed specification (e.g. ¶000202) regarding backbone and head portions of the network. As such, Wu more clearly teaches and/or suggests “wherein the tensor is the sole dividing point of a network that is split into a first portion and second portion.” [See for e.g. 0053 with reference to fig. 4] Given Wu’s teachings above, it would have therefore been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify teachings of Kim and Rao, to add the
techniques of Wu as above for performing device localization and tracking that can accurately estimate both the device feature and the physical device using MR images for interventional procedures (e.g. 0008).
Regarding claim 6, Kim and Rao teach and/or suggest all the limitations of claim 1, and are analyzed as previously discussed with respect to that claim. Kim and Rao however do not address the features of claim 6. On the other hand, the work of Wu from the same or similar field of endeavor is brought in to teach and/or suggest “wherein the first portion and second portion do not include a hierarchical representation of an input to the first portion.” [See for e.g. 0053 with reference to fig. 4 where the device feature Mask R-CNN includes, a backbone architecture 430, a head architecture 432 and a region proposal network (RPN) 434, i.e. it does not include a hierarchical representation (e.g. FPN) of an input to the first portion] The motivation for combining Kim, Rao, and Wu has been discussed in connection with claim 5, above.
Allowable Subject Matter
7. Claims 3 and 9-11 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. In light of the specification, the Examiner finds the claimed invention to be patentably distinct from the prior art of records. The prior art of record, taken individually or in combination fail to explicitly teach or render obvious within the context of the respective independent claims the limitations:
3. The method of claim 1, wherein one or more groups of coefficients are selected based on a distance metric between the tensor and projections of the one or more groups of coefficients with the plurality of basis vectors; and encoding portions of the feature frame bitstream corresponding to the selected one or more groups of coefficients.
9. (Original) The method according to claim 8, wherein the assignment provides an indication of which basis vectors of a plurality of basis vectors are to have corresponding coefficients in a feature frame, and the decoding of the feature frame uses the indication.
10. (Original) The method according to claim 8, wherein the independently coded regions are ordered based on explained variance of the basis vectors in each region.
11. (Original) The method according to claim 8, wherein the indication results in the use of a contiguous subset of the one or more coded regions including the coded region with basis vectors having the highest explained variance.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Please see PTO 892 for additional references.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to RICHARD A HANSELL JR. whose telephone number is (571)270-0615. The examiner can normally be reached Mon - Fri 10 am- 7 pm.
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 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.
/RICHARD A HANSELL JR./Primary Examiner, Art Unit 2486
