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 Amendment
This action is in response to the amendment filed on 1/22/2026. Claims 1, 2, 6, 7, 10, 15, 16, 17, 18, 19, and 20 have been amended, claims 8-9 have been cancelled, and claims 21-22 and have been added. Amendments have overcome the 101 rejection, but fail to overcome the 103 rejections for claims 1, 2, 6, 7, 10, 15, 16, 17, 18, 19, and 20. Claims 21-22 are objected to.
Response to Arguments
In response to applicant’s arguments regarding Ripple nor Mammou failing to teach the newly amended claim 1, Examiner agrees and the office action has been updated to reflect the new limitations of claim 1 by citing Wang as an additional reference for the 103 rejection.
In response to applicant’s argument’s regarding dependent claims being allowable, claims 21-22 are objected to and would be allowable if placed into independent claim form.
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, 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, 3, 4, 5, 6, 7, 10, 11, 12, 13, 15, 17, 18, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Ripple et al. (U.S. Patent Publication No. 2018/0176576), in view of Mammou et al. (U.S. Patent Publication No. 2019/0080483), in further view of Wang et al. (WO 2018/200316).
Regarding claim 1, Ripple discloses determining probabilities for entropy coding of information associated with a current node of the tree, including: selecting a neural network, from two or more pretrained neural networks, according to a level of the current node of the tree (interpreted as for the node currently being coded, choose one model from two or more pretrained neural networks, and the choice is made by the nodes tree level (such as depth))(Ripple: Abstract “the DLBC system applies a first model that is trained to predict feature probabilities based on the context of each bit of the binary codes . The DLBC system compresses the binary code via adaptive arithmetic coding based on the determined probability of each bit.”) obtaining the probabilities through processing input data related to the current node by the selected neural network (interpreted as feed node related inputs into the selected neural network and produce probability values)(Ripple: Abstract “applies a first model that is trained to predict feature probabilities based on the context of each bit of the binary codes”)[Ripple: 0007 “predicts the probability of each bit based on the predicted feature probabilities of the model”](teaches trained model ingests context features and outputs probabilities); and entropy coding of the information associated with the current node based on the probabilities (interpreted as perform an entropy coder that uses those probabilities to code the nodes symbols)(Ripple: Abstract “The DLBC system compresses the binary code via adaptive arithmetic coding based on the determined prob ability of each bit”)(teaches probability driven arithmetic coding corresponding to entropy coding), but fails to explicitly teach a point cloud data coding method applied to an electronic device, the method comprising: obtaining an N-ary tree representation of point cloud data, network, including: determining one or more features for the current node using one or more Multi Layer Perceptron (MLP) network(s),classifying the one or more features into the probabilities of the information associated with the current node of the tree by the one or more MLP network(s), which further includes: applying a multi-dimensional softmax layer and obtaining the probabilities as an output of the multi-dimensional softmax layer.
However, Mammou discloses a point cloud data coding method applied to an electronic device, the method comprising: obtaining an N-ary tree representation of point cloud data (interpreted as having the point cloud expressed as a tree with branching factor N)[Mammou: 0020 “The occupancy symbols are then used by an octree decoding element to recreate a point cloud geometry of the point cloud being decompressed .”][Mammou: 0144 “In some embodiments , the compressed spatial information may have been compressed using an Octree technique and an Octree decoding technique may be used to generate decompressed spatial information for the point cloud .”](teaches octree based representation/processing of point cloud geometry).
However, Wang discloses network, including: determining one or more features for the current node using one or more Multi Layer Perceptron (MLP) network(s), classifying the one or more features into the probabilities of the information associated with the current node of the tree by the one or more MLP network(s), [Wang: 0032 “After receiving the input 210 represented with the octree, convolutional neural network operations can be performed for the nodes of the octree from the bottom up. As shown in FIG. 2, the depth of the neural network is shown in a dashed box 260. In the example architecture 200, a feature map 220 is obtained by performing a convolutional operation on nodes with d-depth in the input 210.”] which further includes: applying a multi-dimensional softmax layer and obtaining the probabilities as an output of the multi-dimensional softmax layer [Wang: 0076 “a softmax layer and two Dropout layers can be added after O-CNN(d), namely”].
Ripple, Mammou, and Wang are considered to be analogous to the claimed invention because they are in the same field of probability driven, point cloud compression. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Ripple to incorporate Mammou and Wang’s teachings of utilizing point clouds as a tree with branching factor N, MLP networks, and softmax layers. The motivation for such a combination would provide the benefit of improved compression efficiency.
Regarding claim 3, Ripple, Mammou and Wang disclose the method according to claim 1, wherein the neural network includes two or more cascaded subnetworks (Ripple: 810-816; Fig. 8)(teaches 2 trained models used sequentially in the same compression pipeline and which are broadly interpreted as two cascade subnetworks).
Regarding claim 4, Ripple and Mammou disclose the method according to claim 1, but fail to explicitly disclose wherein the processing the input data related to the current node comprises: inputting context information for the current node and/or context information for parental and/or neighboring nodes of the current node to a first subnetwork, wherein the context information comprises spatial and/or semantic information.
However, Wang discloses wherein the processing the input data related to the current node comprises: inputting context information for the current node and/or context information for parental and/or neighboring nodes of the current node to a first subnetwork (interpreted as a network stage receives inputs that include features of the current node and optionally its parents and or neighbor nodes, first subnetwork is interpreted as the first stage/module of the NN pipeline) [Wang: 0087 “determining eight nodes having a same parent node from the nodes with the depth associated with the convolutional layer; determining a node neighborhood of the eight nodes based on the octree, the node neighborhood including nodes adjacent to at least one of the eight nodes; and performing the operation of the convolutional layer for the eight nodes on the node neighborhood”][Wang: 0077 “the deconvolutional network can be cascaded with OCNN(d) to obtain the segmentation label on each leaf node”](teaches performing the CNN operation on the node neighborhood built from nodes that share a parent and are adjacent meaning the first CNN stage ingests context from the current nodes and their parent/neighbor nodes. Further teaches cascade networks followed by deconvolutional network, so the front O-CNN corresponds to a first subnetwork receiving that context), wherein the context information comprises spatial and/or semantic information (spatial is interpreted as geometry and semantic is interpreted as attributes or labels like empty/non empty)[Wang: 0085 “obtaining, based on a spatial segmentation for the three-dimensional shape, at least one of: an index of a node of the octree; a label of a node of the octree, the label at least indicating whether the node is an empty node or a non-empty node; shape data of a node of the octree, the shape data at least indicating a shape of the three dimensional shape in the node; and a hierarchical structure of the octree.”](teaches spatial inputs and semantic inputs (labels)).
Ripple, Mammou, and Wang are considered to be analogous to the claimed invention because they are in the same field of probability driven, point cloud compression. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Ripple and Mammou to incorporate Wang’s teachings of utilizing spatial location. The motivation for such a combination would provide the benefit of improved compression efficiency.
Regarding claim 5, Ripple and Wang disclose the method according to claim 4, but fail to explicitly disclose wherein the spatial information includes spatial location information; and wherein the semantic information includes one or more of parent occupancy, tree level, an occupancy pattern of a subset of spatially neighboring nodes, and octant information.
However, Mammou discloses wherein the spatial information includes spatial location information (Mammou: 108; Fig. 1A)(explicit teachings of spatial location); and wherein the semantic information (interpreted a non-geometric descriptors about the nodes state) includes one or more of parent occupancy, tree level (interpreted as node depth) [Mammou: 0291 “At 1288 occupancy symbols for a level of an octree of a point cloud are determined . At 1290 an adaptive look - ahead table with “ N ” symbols is initialized . At 1292 , a cache with “ M ” symbols is initialized . At 1294 , symbols for the current octree level are encoded using the techniques described above . At 1296 , it is determined if additional octree levels are to be encoded.”] (teaches octree level corresponding to tree level), an occupancy pattern of a subset of spatially neighboring nodes (interpreted as neighbor occupancy bits) [Mammou: 0229 “compute the 6 - neighbors used for arithmetic context selection”] [Mammou: 0274 “the occupancy information for the 6 neighbors is obtained by fetching the information directly from the look up table”], and octant information [Mammou: 0016 “To perform the octree encoding , the geometry encoder is configured to use a binary arithmetic encoder to encode up to 256 occupancy symbols . The occupancy symbols represent whether each of 8 sub - cubes of a cube are occupied with points or are un - occupied . Since there are 8 sub - cubes and two - possibilities ( occupied or un - occupied ) , the number of occupancy symbols possible is 28 or 256 .”](teaches multiple semantics, tree levels, neighbor subset occupancy pattern via 6 neighbor occupancy used for context selection, and octant information via the 8 child sub-cubes (octants)).
Ripple, Mammou, and Wang are considered to be analogous to the claimed invention because they are in the same field of probability driven, point cloud compression. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Ripple and Mammou to incorporate Wang’s teachings of utilizing spatial location. The motivation for such a combination would provide the benefit of improved compression efficiency.
Regarding claim 6, Ripple and Mammou disclose the method according to claim 4, but fail to explicitly disclose further comprising determining the one or more features for the current node using the context information as an input to a second subnetwork.
However, Wang discloses further comprising determining the one or more features for the current node using the context information as an input to a second subnetwork [Wang: 0077 “the deconvolutional network can be cascaded with OCNN(d) to obtain the segmentation label on each leaf node. The deconvolutional network is a mirror layout of O-CNN( d), where the deconvolutional and unpooling operations replace the convolutional and pooling operations.”][Wang: 0090 “determining a node neighborhood of the eight nodes based on the octree, the node neighborhood including nodes adjacent to at least one of the eight nodes; and performing an operation of the deconvoluitonal layer for the eight nodes on the node neighborhood.”][Wang: 0075 “Since the feature map is only associated with nodes, its storage space is in linear relation with the number of nodes at the corresponding depth.”](teaches a second subnetwork (the deconvolutional network) cascaded after the first (O-CNN). That second subnetwork ingests context via the node neighborhood and performs its operations on the node neighborhood and further teaches a feature map is associated with nodes which corresponds to features for the current node).
Ripple, Mammou, and Wang are considered to be analogous to the claimed invention because they are in the same field of probability driven, point cloud compression. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Ripple and Mammou to incorporate Wang’s teachings of utilizing features using context information. The motivation for such a combination would provide the benefit of improved representation and code efficiency.
Regarding claim 7, Ripple and Mammou disclose the method according to claim 6, but fail to explicitly disclose further comprising determining one or more features for the current node using one or more Long Short-Term Memory (LSTM) network(s), or determining one or more features for the current node using one or more Convolutional Neural Network (CNN) network(s).
However, Wang discloses further comprising determining one or more features for the current node using one or more Long Short-Term Memory (LSTM) network(s), or determining one or more features for the current node using one or more Convolutional Neural Network (CNN) network(s) [Wang: 0032 “After receiving the input 210 represented with the octree, convolutional neural network operations can be performed for the nodes of the octree from the bottom up. As shown in FIG. 2, the depth of the neural network is shown in a dashed box 260. In the example architecture 200, a feature map 220 is obtained by performing a convolutional operation on nodes with d-depth in the input 210.”] (teaches using a CNN and only one of these limitations need to be met).
Ripple, Mammou, and Wang are considered to be analogous to the claimed invention because they are in the same field of probability driven, point cloud compression. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Ripple and Mammou to incorporate Wang’s teachings of utilizing convolutional neural networks. The motivation for such a combination would provide the benefit of improved code efficiency and implementation flexibility.
Regarding claim 10, Ripple and Wang disclose the method according to claim 7, but fail to explicitly disclose wherein a symbol associated with the current node is an occupancy code.
However, Mammou discloses wherein a symbol associated with the current node is an occupancy code (Mammou: 1288; Fig. 12F “Determine occupancy symbols for a level of an octree of”).
Ripple, Mammou, and Wang are considered to be analogous to the claimed invention because they are in the same field of probability driven, point cloud compression. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Ripple and Wang to incorporate Mammou’s teachings of utilizing occupancy symbols. The motivation for such a combination would provide the benefit of improving compression efficiency.
Regarding claim 11, Ripple and Wang disclose the method according to claim 1, but fail to explicitly disclose wherein the tree representation includes geometry information.
However, Mammou discloses wherein the tree representation includes geometry information [Mammou: 0016 “To perform the octree encoding , the geometry encoder”].
Ripple, Mammou, and Wang are considered to be analogous to the claimed invention because they are in the same field of probability driven, point cloud compression. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Ripple and Wang to incorporate Mammou’s teachings of utilizing geometry encoder. The motivation for such a combination would provide the benefit of improved coding efficiency.
Regarding claim 12, Ripple and Wang disclose the method according to claim 1, but fail to explicitly disclose wherein octree is used for the tree partitioning based on geometry information.
However, Mammou discloses wherein octree is used for the tree partitioning based on geometry information [Mammou: 0016 “To perform the octree encoding , the geometry encoder is configured to use a binary arithmetic encoder to encode up to 256 occupancy symbols .”].
Ripple, Mammou, and Wang are considered to be analogous to the claimed invention because they are in the same field of probability driven, point cloud compression. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Ripple and Wang to incorporate Mammou’s teachings of utilizing geometry encoder. The motivation for such a combination would provide the benefit of improved coding efficiency.
Regarding claim 13, Ripple and Wang disclose the method according to claim 1, but fail to explicitly disclose wherein any of octree, quadtree and/or binary tree or a combination of thereof is used for the tree partitioning based on geometry information.
However, Mammou discloses wherein any of octree, quadtree and/or binary tree or a combination of thereof is used for the tree partitioning based on geometry information [Mammou: 0016 “the octree encoding , the geometry encoder is configured to use a binary arithmetic encoder”].
Ripple, Mammou, and Wang are considered to be analogous to the claimed invention because they are in the same field of probability driven, point cloud compression. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Ripple and Wang to incorporate Mammou’s teachings of utilizing octree. The motivation for such a combination would provide the benefit of improved coding efficiency.
Regarding claim 15, Ripple, Mammou, and Wang disclose the method according to claim 1, Ripple further discloses wherein the entropy coding of the current node further comprises performing arithmetic entropy coding of the symbol associated with the current node using the probabilities (Ripple: Abstract “The DLBC system compresses the binary code via adaptive arithmetic coding based on the determined probability of each bit”).
Claims 17, 18, and 19 are encoding and decoding claims corresponding to claim 1 without any additional limitations. Thus, claims 17, 18, and 19 are rejected for the same reasons as claim 1 above.
Claims 2 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Ripple et al. (U.S. Patent Publication No. 2018/0176576), in view of Mammou et al. (U.S. Patent Publication No. 2019/0080483), in view of Wang et al. (WO 2018/200316), in further view of Sareen et al. (U.S. Patent Publication No. 2016/0247017).
Regarding claim 2, Ripple, Mammou, and Wang disclose the method according to claim 1, Ripple further discloses wherein the selecting the neural network includes: selecting a first neural network in response to the level of the current node exceeding the predefined threshold (Ripple: Abstract “the DLBC system applies a first model that is trained to predict feature probabilities based on the context of each bit of the binary codes . The DLBC system compresses the binary code via adaptive arithmetic coding based on the determined prob ability of each bit”)(teaches distinct training models selected for probability estimation feeding entropy coding and below, Sareen provides the threshold gating by level):
selecting a second neural network, which is different from the first neural network, in response to the level of the current node not exceeding the predefined threshold (same quote as above), but fail to explicitly disclose comparing of the level of the current node within the tree with a predefined threshold.
However, Sareen discloses comparing of the level of the current node within the tree with a predefined threshold [Sareen: 0333 “If a leaf node (which, in an unmodified octree, would contain only one point from the point cloud of the scanned Subject) has a tree depth less than a specified threshold (by way of example and not by limitation, common values for the aforementioned tree depth threshold are 8,9, and 10, with a larger value denoting higher Surface resolution) and generates at least one triangle as per the marching cubes method described above, the node (and its children fulfilling the same criteria) may be split and re-triangulated with marching cubes.”](teaches performing a comparison between node depth level and a specified threshold to decide processing).
Ripple, Mammou, Wang, and Sareen are considered to be analogous to the claimed invention because they are in the same field of probability driven, point cloud compression. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Ripple, Mammou, and Wang to incorporate Sareen’s teachings of utilizing node level comparison. The motivation for such a combination would provide the benefit of improved performance by reducing modeling error and computation.
Claim 20 is a decoding claim corresponding to claim 2 without any additional limitations. Thus, claim 20 is rejected for the same reasons as claim 2 above.
Claim 14 is rejected under 35 U.S.C. 103 as being unpatentable over Ripple et al. (U.S. Patent Publication No. 2018/0176576), in view of Mammou et al. (U.S. Patent Publication No. 2019/0080483), in view of Wang et al. (WO 2018/200316), in further view of Lasserre et al. (U.S. Patent Publication No. 2020/0413080).
Regarding claim 14, Ripple, Mammou, and Wang disclose the method according to claim 1, but fail to explicitly disclose wherein the selecting the neural network is further based on a predefined number of additional parameters, wherein the additional parameters are signaled in a bitstream.
However, Lasserre discloses wherein the selecting the neural network is further based on a predefined number of additional parameters, wherein the additional parameters are signaled in a bitstream [Lasserre: 0064 “Planarity may be signaled with respect to a volume through a planar mode flag… there may be multiple flags : isZPlanar , is YPlanar , is XPlanar”]][Lasserre: 0066 “a more complex signaling syntax may be used involving multiple flags or a non - binary syntax element”][Lasserre: 0067 “The planar mode flag and / or the plane position flag may be encoded in the bitstream”](teaches parameters signaled in the bitstream (flags), these are additional parameters available to drive decisions).
Ripple, Mammou, Wang and Lasserre are considered to be analogous to the claimed invention because they are in the same field of probability driven, point cloud compression. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Ripple, Mammou, and Wang to incorporate Lasserre’s teachings of utilizing additional parameters. The motivation for such a combination would provide the benefit of improving rate distortion and clean encoder/decoder synchronization.
Claim 16 is rejected under 35 U.S.C. 103 as being unpatentable over Ripple et al. (U.S. Patent Publication No. 2018/0176576), in view of Mammou et al. (U.S. Patent Publication No. 2019/0080483), in view of Wang et al. (WO 2018/200316), in further view of Gladding et al. (U.S. Patent Publication No. 2020/0413106).
Regarding claim 16, Ripple, Mammou, and Wang disclose the method according to claim 1, but fail to explicitly disclose wherein the entropy coding of the current node further comprises performing asymmetric numeral systems (ANS) entropy coding of the symbol associated with the current node using the predicted probabilities.
However, Gladding discloses wherein the entropy coding of the current node further comprises performing asymmetric numeral systems (ANS) entropy coding of the symbol associated with the current node using the predicted probabilities (Gladding: Fig. 1 “range asymmetric number system ( “ RANS ” ) encoding and / or RANS decoding”)(teaches ANS).
Ripple, Mammou, Wang, and Gladding are considered to be analogous to the claimed invention because they are in the same field of probability driven, point cloud compression. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Ripple, Mammou, and Wang to incorporate Gladding’s teachings of utilizing ANS. The motivation for such a combination would provide the benefit of improving operation with probability information.
Allowable Subject Matter
Claims 21-22 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
THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). 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 AHMED TAHA whose telephone number is (571)272-6805. The examiner can normally be reached 8:30 am - 5 pm, Mon - Fri. 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, XIAO WU can be reached at (571)272-7761. 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.
/AHMED TAHA/Examiner, Art Unit 2613
/XIAO M WU/Supervisory Patent Examiner, Art Unit 2613
Prosecution Insights