DETAILED ACTION
This action is responsive to the Amendments and Remarks received 02/26/2026 in which claims 2–6, 9, 11–14, 18, 21, 22, 24, and 25 are cancelled, claims 1, 17, 20, and 23 are amended, and no claims are added as new claims.
Response to Arguments
Examiner incorporates herein previous Reponses to Arguments.
On pages 9–14 of the Remarks, Applicant contends the prior art is deficient for failing to teach or suggest the last two “when” clauses of claim 1. Examiner disagrees. Regarding the first “when” clause, Examiner notes Applicant has argued that support for this feature is found in the second priority document. Applicant’s penultimate “when” clause is supported by this disclosure because although the claimed feature requires all three transform skip flags to be false as recited in the claim (set off by “and,” rather than “or”), the disclosed logic in the second priority document is met because the disclosed logic only requires one of the flags to be false such that all flags being false also meets the constraint. Having all flags indicate false as claimed (i.e. transform not skipped for any component) meets the requirement that at least one flag be false as disclosed in Applicant’s priority document. If it’s good enough for the priority document, as averred by Applicant through Applicant’s admissions on the record (see priority discussion, infra), then it’s good enough when the prior art teaches the same thing. Because Applicant’s argument regarding digital logic fails to account for the reality that all flags meeting a requirement fulfills a condition that at least one flag meet a requirement, Applicant’s argument is unpersuasive of error. Applicant’s claims say, “when,” not “only when.” Besides, Egilmez would be interpreted more broadly by the skilled artisan to cover similar scenarios in light of Egilmez’s ¶ 0123, which explains the point of the teachings of Egilmez is to include the possibility that transform skip flags other than solely the luma transform skip flag can inform the signaling of LFNST across the anticipated scenarios. And logically, making the condition require all three skip flags be disabled is the same as the prior art way of just looking at the luma transform skip flag because regardless of the chroma flags, the luma flag controls and must meet the condition (prior to Egilmez, as Egilmez explains, the luma flag essentially controlled (matched) the other flags). Regarding the second “when” clause, Examiner notes this feature is only supported, according to Applicant’s own admission, by Applicant’s third priority filing. This feature is taught by the literal language of Egilmez’s ¶ 0130. Applicant’s argument attempts (unsuccessfully) to reach down further in the paragraph and assign an additional condition to the literal teaching such that Applicant would like the Office to ignore the literal language in favor of constraining the teaching to an example drawn to dual tree. Examiner is not persuaded. Egilmez’s teachings are not constrained to only dual tree. As Egilmez’s ¶ 0123 explains, the point of the teachings of Egilmez is to include the possibility that transform skip flags other than the luma transform skip flag can inform the signaling of LFNST for chroma blocks, not just for dual tree, but especially for dual tree. One skilled in the art would not interpret Egilmez’s teachings to be constrained to a single example provided in Egilmez’s paragraph [0130] drawn to dual tree, but would instead take at face value the broader teachings that begin paragraph [0130] of the prior art reference, especially in light of Egilmez’s explanation that dual tree is a particularly salient case of when it might be desirable to decouple LFNST for all color components from just the luma transform skip flag and instead add more granular control based on treeType and color component-specific transform skip flags. Because the literal language of Egilmez teaches the features of the second recited “when” clause, because Egilmez is clear that the teachings are exemplary regarding dual tree, because the skilled artisan would be able to extend Egilmez’s teachings to other scenarios including single tree, because, in the case of at least the first recited “when” clause, digital logic rules contradict the constrained interpretation averred by Applicant, and because, in the case of at least the second recited “when” clause, Egilmez’s overarching teaching in paragraph [0123] would teach or at least suggest to the skilled artisan that only if all three color components are skipped would there be no need for an LFNST index, Applicant’s arguments are unpersuasive of error.
Other claims are not argued separately. Remarks, 14.
Priority
Examiner incorporates herein previous findings regarding priority.
Examiner notes with appreciation Applicant’s assistance in identifying which priority documents support the claimed features. See interview summary dated 01/30/2025 and documentation provided by Applicant. Examiner notes there are currently no claims supported by the two earliest claimed foreign priority dates.
Applicant claims priority to a foreign filing having a filing date of December 24, 2019. That filing has figures 17 and 18 at the end of the document disclosing pseudocode that supports a decision to decode an LFNST index (lfnst_idx) if certain conditions are met. Specifically the line that states ((treeType == SINGLE_TREE && ( transform_skip_flag[ x0 ][ y0 ][ 0 ] == 0 || transform_skip_flag[ x0 ][ y0 ][ 1 ] == 0 || transform_skip_flag[ x0 ][ y0 ][ 2 ] == 0 )) means the following: (1) if the treeType is single tree and at least one of the transform skip flags for luma or Cb or Cr is false, then the LFNST index is decoded. Applicant’s penultimate “when” clause is supported by this disclosure because although it requires all three transform skip flags to be false as recited in the claim (set off by “and,” rather than “or”), the disclosed logic is met because the disclosed logic only requires one of the flags to be false such that all flags being false also meets the constraint. Having all flags indicate false (i.e. transform not skipped for any component) meets the requirement that at least one flag be false. However, Applicant’s last “when” clause does not appear to be supported by this disclosure because it only requires one component’s transform skip flag to be true as recited in the claim (set off by “or,” rather than “and”) whereas the disclosed logic requires all three transform skip flags to be true to avoid decoding the LFNST index. The last “when” clause, then, requiring only one of the three transform skip flags to be true does not match the requirement of the disclosed logic that requires all three transform skip flags to be true to avoid decoding the LFNST index. This is a critical error in this case because Applicant has asserted this described logic to support Applicant’s claim to foreign priority. See Applicant’s helpful claim chart provided in the record on January 30, 2025 wherein Applicant cites for claims 12 and 13 the same logic. According to this chart, it appears Applicant admits only the third priority document has the requisite support for the last “when” clause of claim 1, as described by Applicant with respect to claim 14 in the provided chart.
Claim Interpretation
Examiner incorporates herein previous findings regarding claim interpretation.
Claim Rejections - 35 USC § 103
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 of this title, 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, 7, 17, 19, 20, 23, and 26 are rejected under 35 U.S.C. 103 as being unpatentable over Koo (US 2021/0360247 A1), Bross et al., “Versatile Video Coding (Draft 7),” JVET-P2001-vE, 16th Meeting: Geneva, CH, Oct. 2019 (uploaded 11/14/2019) (herein “Bross”), and Egilmez (US 2021/0195222 A1).
Regarding claim 1, the combination of Koo, Bross, and Egilmez teaches or suggests a decoding method, comprising: determining an inverse transform method for a target block; and performing an inverse transform for the target block using the inverse transform method (Koo, ¶¶ 0021 and 0023: teaches an inverse transform applied to a target block based on coding parameters such as size and intra prediction mode wherein the transform type/size is determined by indicating which transform in a set of transforms to use and wherein the transform type can be determined by signaling an index value; Koo, ¶ 0167: teaches the transform combination candidates can be selected or determined using indexes; see also Koo, ¶ 0269: teaching NSST or RST indexes), wherein the inverse transform comprises a secondary inverse transform and a primary inverse transform (Koo, Fig. 5: illustrates the inverse transform unit includes both inverse secondary transforms and inverse primary transforms), a secondary inverse transform method of the secondary inverse transform is determined based on a coding parameter for the target block (Koo, ¶¶ 0021 and 0023: teaches an inverse transform applied to a target block based on coding parameters such as size and intra prediction mode), the secondary inverse transform method of the secondary inverse transform is determined based on whether a type of a tree structure for the target block is a dual tree type (Bross, page 71: teaches LFNST is conditioned on treeType being SINGLE_TREE or DUAL_TREE_CHROMA wherein, for example, the LFNST width and height are determined based on dual tree chroma; Examiner notes the algorithm description document related to Bross (JVET-P2002) explains, “If dual tree is enabled, LFNST indices for luma and chroma are signaled separately.” Koo, ¶¶ 0021 and 0023: teaches an inverse transform applied to a target block based on coding parameters such as size and intra prediction mode; see also Koo, ¶ 0110: teaching the transform configuration can be based on block size and block shape; Koo, ¶¶ 0143–0146: teaches different inverse secondary transforms can be applied based on the size and shape; Koo, ¶¶ 0575–0579: teach the size and shape of the block can determine the secondary transform; Koo, ¶¶ 0031, 0079, and 0082: teaches that tree information determines the size and shape of blocks), the secondary inverse transform method is one of multiple methods, a secondary inverse transform method index indicates the secondary inverse transform method (Koo, ¶ 0167: teaches the transform combination candidates can be selected or determined using indexes; see also Koo, ¶ 0269: teaching NSST or RST indexes; Koo, ¶ 0143: teaches the secondary transforms can be MTS transforms indicated by an index specifying, for example, DST 7, DCT 8, etc.; see also Egilmez, ¶¶ 0022 and 0136: teaching the decoding of lfnst_idx, which is a secondary transform index indicating secondary inverse transform, based on tree type and transform skip flags), whether the secondary inverse transform method index is decoded from a bitstream is determined based on the type of the tree structure for the target block (Egilmez, ¶¶ 0022 and 0136: teaching the decoding of lfnst_idx, which is a secondary transform index indicating secondary inverse transform, based on tree type and transform skip flags), a first transform skip mode flag for a Luma component indicates whether a transform skip for the Luma component is used or not, a second transform skip mode flag for a Cb component indicates whether a transform skip for the Cb component is used or not, a third transform skip mode flag for a Cr component indicates whether a transform skip for the Cr component is used or not (Egilmez, ¶ 0129: teaches there are transform skip flags for luma, Cb, and Cr; Examiner notes that the syntax, transform_skip_flag[x0][y0][cIdx] utilizes the color component index, cIdx, to signal luma (0), Cb (1), or Cr (2)), when the type of the tree structure for the target block is a single tree type, the secondary inverse transform method index is decoded in a case that the first transform skip mode flag indicates that a first transform for the Luma component is not skipped, the second transform skip mode flag indicates that a second transform for the Cb component is not skipped and the third transform skip mode flag indicates that a third transform for the Cr component is not skipped (Egilmez, ¶¶ 0129 and 0130: teaches there are transform skip flags for luma, Cb, and Cr (see supra) and that for single tree all three transform skips being disabled would mean the LFNST index would be signaled and decoded), and when the type of the tree structure for the target block is the single tree type, the secondary inverse transform method index is not decoded in a case that the first transform skip mode flag indicates that the first transform for the Luma component is skipped, the second transform skip mode flag indicates that the second transform for the Cb component is skipped, or the third transform skip mode flag indicates that the third transform for the Cr component is skipped (Egilmez, ¶ 0130: teaches there are transform skip flags for luma, Cb, and Cr (see supra) and that for single tree, only if all three transform skips are disabled, then LFNST index would be signaled and decoded, meaning that if any of the three components are transform skipped (transform skip enabled), then LFNST would not be signaled and decoded; Examiner notes Egilmez’s teachings are not constrained to only dual tree; As Egilmez’s ¶ 0123 explains, the point of the teachings of Egilmez is to include the possibility that transform skip flags other than the luma transform skip flag can inform the signaling of LFNST for chroma blocks, not just for dual tree, but especially for dual tree; One skilled in the art would not be constrained to the single example provided in Egilmez’s paragraph [0130], but would instead take at face value the broader teachings that begin that paragraph of the prior art reference and would further be informed by the overarching point articulated in Egilmez’s ¶ 0123 regarding the decoupling of the luma transform skip flag from controlling all LFNST for all color components; In other words, Egilmez’s ¶ 0123 teaches or suggests that just because the luma transform skip flag is true does not mean that LFNST is not still possible for the Cb and Cr chroma components and thus Applicant’s recited requirement that only when it is assured all three color components are skipped (set-off by AND logic), then LFNST safe to exclude from the bitstream is obvious).
One of ordinary skill in the art, before the effective filing date of the claimed invention, would have been motivated to combine the elements taught by Koo, with those of Bross, because both references are drawn to the same field of endeavor such that one wishing to implement secondary transforms would be led to their relevant teachings and because combining Koo’s explanations regarding intra-prediction mode or other parameters informing secondary transforms and the ability to further split into subblocks (e.g. Koo, ¶ 0129) with Bross explaining LFNST can be informed by treeType represents a mere combination of prior art elements, according to known methods, to yield a predictable result. This rationale applies to all combinations of Koo and Bross used in this Office Action unless otherwise noted.
One of ordinary skill in the art, before the effective filing date of the claimed invention, would have been motivated to combine the elements taught by Koo and Bross, with those of Egilmez, because all three references are drawn to the same field of endeavor such that one wishing to implement secondary transforms would be led to their relevant teachings and because combining Koo’s explanations regarding intra-prediction mode or other parameters informing secondary transforms and the ability to further split into subblocks (e.g. Koo, ¶ 0129) with Bross’s and Egilmez’s explanation that LFNST can be informed by treeType represents a mere combination of prior art elements, according to known methods, to yield a predictable result. This rationale applies to all combinations of Koo, Bross, and Egilmez used in this Office Action unless otherwise noted.
Regarding claim 7, the combination of Koo, Bross, and Egilmez teaches or suggests the decoding method of claim 1, wherein the secondary inverse transform method index is included in a bitstream when the coding parameter has a specific value (Koo, ¶ 0140: teaches the secondary transform may be conditionally applied such that a coding parameter, such as whether the block is an intra block, can control whether secondary transform is enabled).
Claim 17 lists at least the elements of claim 1 but is drawn to an encoding method instead of the corresponding decoding method. Therefore, the rationale for the rejection of claim 1 applies to the instant claim.
Claim 19 lists the same elements as claim 17, but is drawn to a product-by-process rather than the method. Therefore, the rationale for the rejection of claim 17 applies to the instant claim.
Claim 20 lists the same elements as claim 17, but is drawn to a product-by-process rather than the method. Therefore, the rationale for the rejection of claim 17 applies to the instant claim.
Claim 23 lists at least the elements of claim 7, but is drawn to an encoding method instead of the corresponding decoding method. Therefore, the rationale for the rejection of claim 7 applies to the instant claim.
Claim 26 lists at least the elements of claim 7, but is drawn to a product-by-process rather than the method. Therefore, the rationale for the rejection of claim 7 applies to the instant claim.
Claims 8, 10, 15, and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Koo, Bross, Egilmez, Zhao (US 2020/0374531 A1) (herein “Zhao ‘531”) and Han (US 2017/0078672 A1).
Regarding claim 8, the combination of Koo, Bross, Egilmez, Zhao ‘531, and Han teaches or suggests the decoding method of claim 1, wherein the secondary inverse transform method is applied in common to a plurality of subblocks in a case that the target block is partitioned into the plurality of subblocks as an intra-sub-partitioning flag indicates that intra-sub-partitioning is applied to the target block (Zhao ‘531, ¶ 0120: teaches the “sub-blocks share a same type of…secondary transform” based on the signaled index value; Bross, page 71: teaches LFNST is conditioned on IntraSubPartitionsSplitType; Han, ¶ 0055: teaches using a same transform type for all sub-blocks of a partitioned block because it is more efficient).
One of ordinary skill in the art, before the effective filing date of the claimed invention, would have been motivated to combine the elements taught by Koo, Bross, and Egilmez, with those of Zhao ‘531, because all four references are drawn to the same field of endeavor such that one wishing to implement secondary transforms would be led to their relevant teachings and because combining (1) Koo’s explanations regarding intra-prediction mode informing secondary transforms and the ability to further split into subblocks (e.g. Koo, ¶ 0129) with Zhao ‘531 explaining the ubiquity of treating like subblocks alike; and (2) Bross’s and Egilmez’s explanations regarding the interaction between ISP and LFNST with Koo and Zhao’ 531 evidences the already-known interplay between those coding tools such that the combination represents a mere combination of prior art elements, according to known methods, to yield a predictable result. This rationale applies to all combinations of Koo, Bross, Egilmez, and Zhao ‘531 used in this Office Action unless otherwise noted.
One of ordinary skill in the art, before the effective filing date of the claimed invention, would have been motivated to combine the elements taught by Koo, Bross, Egilmez, and Zhao ‘531, with those of Han, because all five references are drawn to the same field of endeavor such that one wishing to implement frequency transforms for video coding would be led to their relevant teachings, because Bross’s and Egilmez’s teachings regarding ISP evidences the already-known interplay between ISP and LFNST, and because, as Han explains, using the same transform type for Bross’s sub-partitions is a known approach for reducing complexity. Therefore, the combination represents a mere combination of prior art elements, according to known methods, to yield a predictable result. This rationale applies to all combinations of Koo, Bross, Egilmez, Zhao ‘531, and Han used in this Office Action unless otherwise noted.
Regarding claim 10, the combination of Koo, Bross, Egilmez, Zhao ‘531, and Han teaches or suggests the decoding method of claim 9, wherein the secondary inverse transform method index is derived as a first value indicating that the secondary inverse transform is not applied in a case that the secondary inverse transform method index is not signaled via a bitstream (Zhao ‘531, ¶ 0132: teaches the secondary transform index value is set to zero, indicating no secondary transform, when the secondary transform index is not signaled in the bitstream).
Regarding claim 15, the combination of Koo, Bross, Egilmez, Zhao ‘531, and Han teaches or suggests the decoding method of claim 1, wherein the secondary inverse transform method index is decoded in a case that a type of the tree structure for the target block is a dual tree luma type and a first transform skip mode flag for a Luma component indicates that a transform skip for the Luma component is not used (Examiner notes the skilled artisan understands transform skip is not compatible with LFNST since LFNST is a transform; The fact that no skip mode flag indicates a transform is skipped for the luma component allows the coding logic to determine whether a secondary transform is applied to the luma component and if so what transform to use; Zhao ‘531, ¶ 0133: teaches, “NSST may not be applied for a block coded with transform skip mode.”; see also Bross, page 71).
Regarding claim 16, the combination of Koo, Bross, Egilmez, Zhao ‘531, and Han teaches or suggests the decoding method of claim 1, wherein the secondary inverse transform method index is decoded in a case that a type of the tree structure for the target block is a dual tree chroma type, a second transform skip mode flag for a Cb component indicates that a transform skip for the Cb component is not used and a third transform skip mode flag for a Cr component indicates that a transform skip for the Cr component is not used (Examiner notes the skilled artisan understands transform skip is not compatible with LFNST since LFNST is a transform; The fact that no skip mode flags indicate a transform is skipped for the chroma components allows the coding logic to determine whether a secondary transform is applied to the chroma components and if so what transform to use; Zhao ‘531, ¶ 0133: teaches, “NSST may not be applied for a block coded with transform skip mode.”; see also Bross, page 71).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Koo (US 2022/0394300 A1) teaches LFNST is a type of secondary transform.
Zhao (US 2020/0374531 A1) teaches when RST index is zero, RST is not applied and also teaches the RST index chooses one of several transform kernels (¶ 0141).
Nalci (US 2020/0396487 A1) teaches RST, NSST, “secondary transform,” and LFNST all “have the same meaning.” (¶ 0086).
Ma et al., “On LFNST signaling with Transform-skip mode,” JVET-Q0499, 17th Meeting: Brussels, BE Jan 2020 (uploaded 01/01/2020).
Yoo et al., “CE8-related: Unified condition between BDPCM and transform skip,” JVET-O0206_v1, 15th Meeting: Gothenburg, SE, July 2019. This publication was applied to claim 11. Examiner notes the skilled artisan understands BDPCM is a spatial domain coding tool and is not compatible with secondary transforms since transforms convert the data into the frequency domain; Yoo, Abstract: explains BDPCM is only used when transform is skipped; In view of this teaching, Examiner finds obviously the secondary transform signaling is influenced by whether BDPCM is being used as the two are not compatible. Yoo merely explains what the skilled artisan already knows regarding how secondary transforms are not compatible with BDPCM.
Han (US 2017/0078672 A1) teaches using a same transform type for all sub-blocks of a partitioned block because it is more efficient (¶ 0055).
Lee (US 2019/0281297 A1) teaches all sub-blocks may use the same transform scheme (¶ 0191).
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 extension fee 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 Michael J Hess whose telephone number is (571)270-7933. The examiner can normally be reached Mon - Fri 9:00am-5:30pm.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, William Vaughn can be reached on (571)272-3922. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8933.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/MICHAEL J HESS/Examiner, Art Unit 2481
Prosecution Insights