13
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 .
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1, 9, and 10 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Regarding claim 1, limitation (a) recites “a data processing step of receiving original 3D data and anchor point number information, and acquiring a number of sampled anchor points corresponding to 3D data subsequent to processing and the anchor point number information based on the original 3D data and anchor point number information.” The phrase uses "anchor point number information" twice in quick succession. This makes the sentence circular and introduces ambiguity about whether the two instances refer to the exact same information or two different sets of information. If the two instances refer to the exact same information, then there is not sufficient support from the specification how the anchor point number information can be based on itself. The specification rather details how the anchor point number information is based on the original 3D data, specification para [62]. The exact relationship between the "original 3D data" and the "anchor point number information" is vague.
For the purpose of compact prosecution and art rejection, the examiner will treat limitation (a) to mean “a data processing step of receiving original 3D data and anchor point number information, and acquiring a number of sampled anchor points corresponding to 3D data subsequent to processing and the anchor point number information based on the original 3D data .”
Further, regarding limitation (b) of claim 1, it recites “…acquiring sampled transformation parameters and transformation parameter values from the transformation parameter and transformation parameter category.” The term "transformation parameter" is used in both the source and the outcome, making the distinction between the two unclear. It is circular, implying that something is being attained from itself. This introduces ambiguity about whether the two instances refer to the exact same information or two different sets of information. If the two instances refer to the exact same information, then there is not sufficient support from the specification how the transformation parameters can be sampled from itself.
Additionally, the claim language “…acquiring sampled transformation parameters and transformation parameter values from the transformation parameter and transformation parameter category.” is inconsistent with the specification. The specification suggests that transformation parameter values are not sampled from transformation parameters; they should be sampled from transformation parameter categories (specification para 73; specification Fig. 8). These contradict the claim.
There is a conflict between the claimed subject matter and the specification disclosure. This renders the scope of the claim uncertain as inconsistency with the specification disclosure makes the claim take on an unreasonable degree of uncertainty. Reference can be made to MPEP 2173.03.
For the purpose of compact prosecution and art rejection, the examiner will treat this claim element to mean …acquiring sampled transformation parameters and transformation parameter values from the
Regarding claim 9, it recites a storage that stores large-capacity network data. The term “large-capacity” is a relative term. One having ordinary skill in the art would not be able to determine the metes and bounds of the term large-capacity only based on the claim language. The specification does not provide guidance on how to determine the metes and bounds of the term large-capacity (specification para 18).
Further, claim 9 recites a first operation of receiving original 3D data and anchor point number information, and acquiring a number of sampled anchor points corresponding to 3D data subsequent to processing and the anchor point number information based on the original 3D data and anchor point number information; The phrase uses "anchor point number information" twice in quick succession. This makes the sentence circular and introduces ambiguity about whether the two instances refer to the exact same information or two different sets of information. If the two instances refer to the exact same information, then there is not sufficient support from the specification how the anchor point number information can be based on itself. The specification rather details how the anchor point number information is based on the original 3D data, specification para [62]. The exact relationship between the "original 3D data" and the "anchor point number information" is vague.
For the purpose of compact prosecution and art rejection, the examiner will treat this limitation to mean a first operation of receiving original 3D data and anchor point number information, and acquiring a number of sampled anchor points corresponding to 3D data subsequent to processing and the anchor point number information based on the original 3D data
Further, claim 9 recites a second operation of receiving at least one transformation parameter and transformation parameter category, and acquiring sample transformation parameter values sampled from the transformation parameter and transformation parameter category. The language in claim 9 “acquiring sample transformation parameter values sampled from the transformation parameter and transformation parameter category” is inconsistent with the specification. The specification suggests that sample transformation parameter values are not sampled from transformation parameters; they should be sampled from transformation parameter categories (specification para 73; specification Fig. 8). These contradict the claim.
There is a conflict between the claimed subject matter and the specification disclosure. This renders the scope of the claim uncertain as inconsistency with the specification disclosure makes the claim take on an unreasonable degree of uncertainty. Reference can be made to MPEP 2173.03.
For the purpose of compact prosecution and art rejection, the examiner will treat this claim element to mean … sampled from the .
Regarding claim 10, it is rejected using the same citations and rationales described in the rejection of claim 9.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-10 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1: Claims 1-8 are drawn to a method (i.e., a process), claims 9-10 are drawn to a device (i.e., a manufacture). As such, claims 1-10 are drawn to one of the statutory categories of invention (Step 1: Yes).
Step 2A – Prong 1: Claims 1, 9, and 10 recite:
(a) a data processing step of receiving original 3D data and anchor point number information, and acquiring a number of sampled anchor points corresponding to 3D data subsequent to processing and the anchor point number information based on the original 3D data and anchor point number information;
(b) a transformation parameter sampling step of receiving at least one transformation parameter and transformation parameter category, and acquiring sampled transformation parameters and transformation parameter values from the transformation parameter and transformation parameter category; and
(c) a data augmentation step of calculating a non-rigid transformation matrix based on the 3D data subsequent to processing, sampled anchor points, and sampled transformation parameter values, and acquiring 3D data transformed by using the non-rigid transformation matrix.
Limitations (a)-(b) can practically be performed in the human mind, with or without the use of a physical aid such as pen and paper; therefore the limitation falls within the mental processes grouping, and the claim recites an abstract idea. Limitation (c) is an act of calculating using mathematical methods to determine a variable (transformed 3D data), and is a mathematical calculation. (Step 2A – Prong 1: Yes).
Step 2A – Prong 2: This judicial exception is not integrated into a practical application because the additional elements are no more than insignificant extra-solution activity and no more than mere instructions to apply the exception using a generic computer component.
The claims recite the additional element of receiving original 3D data and anchor point number information, and acquiring a number of sampled anchor points in limitation (a). The additional element represents mere data gathering that is necessary for use of the recited judicial exception and is recited at a high level of generality. Limitation (a) in the claim is thus insignificant extra-solution activity.
Limitations (b) and (c) is an algorithm to sample at least one transformation parameter and transformation parameter category, and calculate a non-rigid transformation matrix and acquire three-dimensional data transformed by using the non-rigid transformation matrix, but the process is recited so generically that it represents no more than mere instructions to apply the judicial exceptions on a computer. Even when viewed in combination, these additional elements do not integrate the recited judicial exception into a practical application and the claim is directed to the judicial exception (Step 2A – Prong 2: No).
Step 2B: The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the processor(s), memory, network interface, and storage are at best the equivalent of merely adding the words "apply it" to the judicial exception. Mere instructions to apply an exception cannot provide an inventive concept. The claims lack affirmative recitation a specific, unconventional technical means. There is no detail that the transformation parameter sampling step, calculation of a non-rigid transformation matrix, or transformation of three-dimensional data by using the non-rigid transformation matrix uses novel algorithms, hardware optimization, or produces a concrete technological improvement (e.g., enhanced memory efficiency, optimized matrix multiplication, reduced bandwidth through a novel data structure). Therefore, when considered separately and in combination, without the aforementioned details, a court is likely to view the steps as routine computer implementation.
Regarding claim 2, it recites: a step (a-1) of performing preprocessing on the original 3D data; and a step (a-2) of sampling a plurality of anchor points that serve as local transformation references on the original 3D data or 3D data subsequent to processing. The limitation “performing preprocessing…” can reasonably be performed by the human mind, and is a mental process or by a human using pen and paper (labeling, segmentation, data cleaning). The limitation “sampling a plurality of anchor points…” can reasonably be performed by the human mind (data organization). Further, the claim does not recite any additional element.
Regarding claim 3, it recites: performing at least one preprocessing of vertex sampling, centralization, or denoising on the original 3D data. The limitation “performing…” can reasonably be performed by the human mind, and is a mental process. Further, the claim does not recite any additional element.
Regarding claim 4, it recites: sampling a first arbitrary anchor point on the original 3D data or 3D data subsequent to processing; and sampling a second anchor point that is present at a position farthest away from the first anchor point on the original 3D data or 3D data subsequent to processing. The limitations “sampling a first arbitrary anchor point…” and “sampling a second anchor point…” are functions that can reasonably be performed by the human mind (data organization), and is a mental process. Further, the claim does not recite any additional element.
Regarding claim 5, it recites: the step (a) is further receiving scene data, wherein the step (a-2) comprises sampling a plurality of anchor points targeting the scene data, the step (a-2) further comprising: segmenting a plurality of instances in the scene data; and sampling one or more anchor points on at least one instance from among the plurality of segmented instances. The limitation “further receiving scene data” is nothing more than an insignificant extra solution activity and not integrated into a practical application under Step 2A Prong 2. Under Step 2B, the limitation “further receiving scene data” is nothing more than an insignificant extra solution activity and considering it both individually and in combination, does not amount to significantly more than the judicial exception itself. The limitation “sampling…” is a function that can reasonably be performed by the human mind, and is a mental process. The limitation “segmenting…” is a function that can reasonably be performed by the human mind (data organization), and is a mental process. Further, the claim does not recite any additional element.
Regarding claim 6, it recites: wherein the transformation parameter comprises at least one of rotation transformation, scaling, and translation. The limitation “wherein the transformation parameter comprises…” does not recite steps of a process. Further, the claim does not recite any additional element.
Regarding claim 7, it recites: wherein the sampled transformation parameter values are random values extracted from within the transformation parameter category. The limitation “wherein the sampled transformation parameter values are…” does not recite steps of a process. Further, the claim does not recite any additional element.
Regarding claim 8, it recites: the step (c) comprises: a step (c-1) of applying transformation parameters extracted from N anchor points to calculate N local transformation matrices; a step (c-2) of linearly combining the N local transformation matrices to acquire a non-rigid transformation matrix; and a step (c-3) of applying the non-rigid transformation matrix to the 3D data subsequent to processing to acquire transformed 3D data. The limitation “applying transformation parameters…” is an act of calculating using mathematical methods to determine variables (local transformation matrices), and is a mathematical calculation. The limitation “linearly combining...” is an act of calculating using mathematical methods to determine a variable (non-rigid transformation matrix), and is a mathematical calculation. The limitation “applying the non-rigid transformation matrix...” is an act of calculating using mathematical methods to determine a variable (transformed 3D data), and is a mathematical calculation. (Step 2B: No).
Further, claim 10 is rejected under 35 U.S.C. 101 because claim 10 recites: “A computer program...” the body of the claim recites computer program steps, such as, “receiving data, sampling at least one transformation parameter and transformation parameter category, and calculate a non-rigid transformation matrix and acquire three-dimensional data transformed by using the non-rigid transformation matrix,” which are nothing more than just programmed instructions to be performed by the system. Therefore, the steps/elements recited in claim 10 are non-statutory. Similarly, computer programs claimed as computer listings per se, ie., the descriptions or expressions of the programs, are not physical “things.” They are neither computer components nor statutory processes, as they are not “acts” being performed. Such claimed computer programs do not define any structural and functional interrelationships between the computer program and other claimed elements of a computer which permit the computer program’s functionality to be realized. In contrast, a claimed non-transitory computer-readable medium encoded with a computer program is a computer element which defines structural and functional interrelationships between the computer program and the rest of the computer which permit the computer program’s functionality to be realized, and is thus statutory. Accordingly, it is important to distinguish claims that define descriptive material per se from claims that define statutory inventions. The examiner notes, the specification does not provide sufficient disclosure that the computer-readable medium is non-transitory and/ or can not be a signal.
Therefore, claims 1-10 are not eligible subject matter under 35 USC 101.
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 1-4 and 6-10 are rejected under 35 U.S.C. 103 as being unpatentable over Juppe (US 20210366205 A1) in view of Kim et al. (Kim et al., "Point Cloud Augmentation with Weighted Local Transformations", 11 Oct. 2021, pages 1-9) (Hereinafter referred to as Kim) in further view of Hacinecipoglu et al. (Hacinecipoglu et al., "Pose Invariant People Detection in Point Clouds for Mobile Robots", 5 May 2020, International Journal of Mechanical Engineering and Robotics Research, Vol. 9, pages 709 - 715) (Hereinafter referred to as Hacinecipoglu) in further view of Zhang (CN 115937458 A).
Regarding claim 9, Juppe teaches a 3D data augmentation apparatus, the apparatus comprising: one or more processors ("In some embodiments of the present technology, the processor may be a general purpose processor, such as a central processing unit (CPU) or a processor dedicated to a specific purpose," para [0062].); a network interface ("The input/output interface 150 may allow enabling networking capabilities such as wire or wireless access. As an example, the input/output interface 150 may comprise a networking interface such as, but not limited to, a network port, a network socket, a network interface controller and the like," para [0082].); a memory that loads a computer program executed by the processor ("According to implementations of the present technology, the solid-state drive 120 stores program instructions suitable for being loaded into the random access memory 130 and executed by the processor 110 for augmenting the 3D objects training dataset," para [0083].); and
a storage that stores large-capacity network data and the computer program ("According to implementations of the present technology, the solid-state drive 120 stores program instructions suitable for being loaded into the random access memory 130 and executed by the processor 110 for augmenting the 3D objects training dataset," para [0083]. The solid-state drive is mapped to the storage. The program instructions read on the computer program.
"The solid-state drive 120 may also store various databases including 3D objects training dataset, metadata, user information such as login, activity history or the like," para [0083]. The 3D objects training dataset and metadata read on the large-capacity network data.),
Juppe fails to teach but Kim teaches wherein the computer program executes, by the one or more processors, a first operation of receiving original 3D data and anchor point number information (Kim; The original point cloud as input reads on receiving original 3D data. (page 4, Algorithm 1 PointWOLF). The “# anchor points M” as input reads on receiving anchor point information (Kim; page 4, Algorithm 1 PointWOLF).), and acquiring a number of sampled anchor points corresponding to 3D data Kim teaches that anchor points are sampled by using the Farthest Point Sampling (FPS) algorithm (page 3, Sampling Anchor Points). FPS uses the original point cloud (mapped to original 3D data) as a parameter (page 4, Algorithm 1 PointWOLF).)
a second operation of receiving at least one transformation parameter (Kim; "S is a diagonal matrix with three positive real values, i.e., S = diag(sx, sy, sz) to allow different scaling factors for different axes." The diagonal matrix S indicates at least one transformation parameter. (pages 3-4, Local Transformations).) and transformation parameter category (Kim; "Input: range for scaling ρs, range for rotation ρr, range for translation ρt," (page 4, Algorithm 1 PointWOLF). The ranges read on at least one transformation parameter category.), and acquiring sample transformation parameter values (Kim; "Local transformations in our framework are centered at the anchor points. At each anchor point, we randomly sample a local transformation that includes scaling from the anchor point, changing aspect ratios, translation, and rotation around the anchor point," (pages 3-4, Local Transformations). The local transformations read on the sample transformation parameter values.) sampled from the transformation parameter and transformation parameter category (For the purpose of compact prosecution and art rejection, the examiner will treat this claim element to mean … sampled from the . (page 4, Algorithm 1 PointWOLF) Lines 3-5 show that the sampled transformation parameters are within the input ranges (which read on transformation parameter category).); and
a third operation of calculating a non-rigid transformation matrix based on the 3D data “Given M local transformations {Tj }M j =1, our smoothly varying transformation at an arbitrary point pi is given as:
PNG
media_image1.png
53
281
media_image1.png
Greyscale
where Kh( · , · ) is a kernel function with bandwidth h, and Tj is the local transformation in (2) centered at pjA . To define
T
^
(pi) at any point in the 3D space, we use a kernel function that has a strictly positive value for any pair of points…”(page 4, Smooth Deformations; page 4, Eq. (3)). The local transformation Tj reads on sampled transformation parameter. Variable pjA (seen in equation 3) is an anchor point in PA, (pages 3-4, Local Transformations); “the anchor points PA ⊂ P are selected by the Farthest Point Sampling (FPS) algorithm,” and thus pjA represents sampled anchor points, which are used in the calculation of the non-rigid transformation matrix. The examiners interpretation is that point pi is an arbitrary point from the 3D data (page 4, Smooth Deformations).
T
^
is used to calculate a non-rigid transformation matrix ), and acquiring 3D data transformed by using the non-rigid transformation matrix(“Thus, our framework can be implemented in two ways: (1) Transforming each point once by a smoothly varying transformation
T
^
in Eq. (3) and (2) Transforming each point M times by the local transformations {Tj }M j =1 and interpolate these M augmented points by the adaptive weights K( p , pjA) / ∑k K( p , pkA ) . Although both approaches re-quire O ( M N ) complexity if we mainly consider M anchor points and N points, the second approach is slightly more efficient in practice since this only involves operations on points (vector) while the first approach involves operations on transformation matrices. Thus, we show the pseudocode of the second approach in Algorithm 1 and show the first approach’s pseudocode in the supplement,”
(page 4, Proposition1); (page 4, Algorithm 1 PointWOLF). Algorithm 1 shows the augmented point cloud as output after applying Eq. 3. The augmented point cloud is mapped to 3D data transformed by using the non-rigid transformation matrix).
Before the effective filling date of the claimed invention, it would have been
obvious to one having ordinary skill in the art to apply the teachings of Kim to Juppe. The motivation would have been to generate “diverse and realistic local deformations such as a person with varying postures,” (Kim; page 2; Introduction) and to maximize the coverage of anchor points and allow diverse transformations (Kim; page 3, Sampling Anchor Points).
Juppe in view of Kim does not explicitly disclose that the 3D data that corresponds to the number of sampled anchor points is subsequent to processing or that the 3D data that calculating a non-rigid transformation matrix is based on is subsequent to processing. Zhang teaches the 3D data is subsequent to processing (“Step 102, preprocessing the acquired point cloud data. In order to improve the modeling accuracy, the method of preprocessing the acquired point cloud data may include filtering, denoising and repairing the acquired point cloud data,” (Detailed Ways, para 3-24). Zhang discloses 3D point cloud data preprocessing. After combination, preprocessing occurs before the main algorithm and the 3D data is utilized subsequent to processing.
Before the effective filling date of the claimed invention, it would have been
obvious to one having ordinary skill in the art to apply the teachings of Zhang to Juppe in view of Kim. The motivation would have been “to improve the modeling accuracy,” (Zhang; Detailed Ways, para 17).
Juppe in view of Kim in further view of Zhang fails to teach but Hacinecipoglu teaches and the anchor point number information based on the original 3D data and anchor point number information (For the purpose of compact prosecution and art rejection, the examiner will treat “and the anchor point number information based on the original 3D data and anchor point number information” to mean and the anchor point number information based on the original 3D data .” Hacinecipoglu; (page 711, Voxel Grid Filtering). Here, the reference controls the number of anchor points (anchor point number information) by changing density of the anchor/representative points. After the combination of Juppe in view of Kim in further view of Hacinecipoglu, one having ordinary skills in the art would determine the M value (number of anchor points) by changing the density of anchor/representative points based on the original point cloud's attributes, e.g., volume, quality, and/or object type.);
Before the effective filling date of the claimed invention, it would have been obvious to one having ordinary skill in the art to apply the teachings of Hacinecipoglu to Juppe in view of Kim in further view of Zhang. The motivation would have been to reduce (downsample) the point cloud “to a reasonable number of points which makes processing easier without losing important features,” (page 711, Voxel Grid Filtering).
Regarding claims 1 and 10, they are rejected using the same citations and
rationales described in the rejection of claim 9.
Regarding claim 2, Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu teaches the method of claim 1, wherein the step (a) comprises: a step (a-1) of performing preprocessing on the original 3D data (Zhang; “Step 102, preprocessing the acquired point cloud data. In order to improve the modeling accuracy, the method of preprocessing the acquired point cloud data may include filtering, denoising and repairing the acquired point cloud data,” (Zhang; Detailed Ways, para 3-24). The acquired point cloud data is mapped to original 3D data); and a step (a-2) of sampling a plurality of anchor points that serve as local transformation references on the original 3D data or 3D data subsequent to processing (Kim; “To minimize the redundancy between local transformations, the anchor points PA ⊂ P are selected by the Farthest Point Sampling (FPS) algorithm.” (page 3, Sampling Anchor points). "Local transformations in our framework are centered at the anchor points. At each anchor point, we randomly sample a local transformation ... Given an anchor point pjA in PA, the local transformation for an input point pi can be written as…" (pages 3-4, Local Transformations). "Figure 2. PointWOLF Framework Illustration. Given an original sample, PointWOLF has multiple local transformations at each anchor point (red). PointWOLF produces smoothly varying non-rigid deformations based on the weighted local transformations," (page 3, Figure 2). The original sample reads on original 3D data.).
Before the effective filling date of the claimed invention, it would have been obvious to one having ordinary skill in the art to modify Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu with the feature of sampling a plurality of anchor points from Kim for the benefit minimizing the redundancy between local transformations (Kim; page 3, Sampling Anchor points).
Regarding claim 3, Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu teaches the method of claim 2, wherein the step (a-1) is performing at least one preprocessing of vertex sampling, centralization, or denoising on the original 3D data (Zhang; “Step 102, preprocessing the acquired point cloud data. In order to improve the modeling accuracy, the method of preprocessing the acquired point cloud data may include filtering, denoising and repairing the acquired point cloud data,” (Zhang; Detailed Ways, para 3-24). The acquired point cloud data is mapped to original 3D data).).
Regarding claim 4, Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu teaches the method of claim 2, wherein the step (a-2) comprises: sampling a first arbitrary anchor point on the original 3D data or 3D data subsequent to processing; and sampling a second anchor point that is present at a position farthest away from the first anchor point on the original 3D data or 3D data subsequent to processing (Kim; “Sampling anchor points is the first step of our framework to locate multiple local transformations. To minimize the redundancy between local transformations, the anchor points PA ⊂ P are selected by the Farthest Point Sampling (FPS) algorithm. FPS randomly chooses the first point and then sequentially chooses the farthest points from previous points. This maximizes the coverage of anchor points and allows diverse transformations,” (page 3, Sampling Anchor Points). “Randomly chooses the first point” reads on “sampling a first arbitrary anchor point.” P refers to the original point cloud (page 4, Algorithm 1 PointWOLF). The original point cloud reads on original 3D data.).
Before the effective filling date of the claimed invention, it would have been obvious to one having ordinary skill in the art to modify Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu with the feature of sampling a first arbitrary anchor point and sampling a second anchor point from Kim for the benefit of minimizing the redundancy between local transformations (Kim; page 3, Sampling Anchor points).
Regarding claim 6, Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu teaches the method of claim 1, wherein the transformation parameter comprises at least one of rotation transformation, scaling, and translation (Kim; "S is a diagonal matrix with three positive real values, i.e., S = diag(sx, sy, sz) to allow different scaling factors for different axes." The diagonal matrix S indicates at least one transformation parameter. (pages 3-4, Local Transformations).)).
Before the effective filling date of the claimed invention, it would have been obvious to one having ordinary skill in the art to modify Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu with the feature of the transformation parameter comprising at least one of rotation transformation, scaling, and translation from Kim for the benefit of expanding and diversifying the original data.
Regarding claim 7, Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu teaches the method of claim 6, wherein the sampled transformation parameter values are random values extracted from within the transformation parameter category (Kim; "Local transformations in our framework are centered at the anchor points. At each anchor point, we randomly sample a local transformation that includes scaling from the anchor point, changing aspect ratios, translation, and rotation around the anchor point," (pages 3-4, Local Transformations). Lines 3-5 of Algorithm 1 show that the transformation parameter values are sampled within the ranges that read on transformation parameter category (Kim; page 4, Algorithm 1 PointWOLF).).
Before the effective filling date of the claimed invention, it would have been obvious to one having ordinary skill in the art to modify Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu with the feature of sampled transformation parameter values being random values extracted from within the transformation parameter category from Kim for the benefit of allowing diverse transformations.
Regarding claim 8, Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu teaches the method of claim 1, wherein the step (c) comprises: a step (c-1) of applying transformation parameters extracted from N anchor points to calculate N local transformation matrices (Kim; "At each anchor point, we randomly sample a local transformation that includes scaling from the anchor point, changing aspect ratios, translation, and rotation around the anchor point," (pages 3-4, Local Transformations). This step provides transformation parameters from the anchor points. “Given an anchor point pjA in PA, the local transformation for an input point pi can be written as:"
PNG
media_image2.png
32
382
media_image2.png
Greyscale
(pages 3-4, Local Transformations; page 4, Eq. 2). This is a one-to-one mapping of an anchor point to a local transformation matrix; this reads on N anchor points to calculate N local transformation matrices.); a step (c-2) of linearly combining the N local transformation matrices to acquire a non-rigid transformation matrix (Kim; “Given M local transformations { Tj } M j =1, our smoothly varying transformation at an arbitrary point pi is given as:
PNG
media_image3.png
51
282
media_image3.png
Greyscale
(page 4, Smooth Deformations; page 4, Eq. 3). M local transformations reads on N local transformation matrices. Eq 3 shows the linear combination of the M local transformations to acquire a non-rigid transformation matrix.); and a step (c-3) of applying the non-rigid transformation matrix to the 3D data subsequent to processing to acquire transformed 3D data (Kim; “Thus, our framework can be implemented in two ways: (1) Transforming each point once by a smoothly varying transformation
T
^
in Eq. (3)…” (page 4, Proposition1; page 4, Algorithm 1 PointWOLF). The “each point” language reads on the 3D data. Zhang teaches the 3D data is subsequent to processing (Detailed Ways, para 3-24).).
Before the effective filling date of the claimed invention, it would have been obvious to one having ordinary skill in the art to modify Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu with the feature of applying transformation parameters, linearly combining the N local transformation matrices, and applying the non-rigid transformation matrix from Kim for the benefit of allowing smooth transformations under mild conditions (Kim; page 4, Smooth Deformations).
Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu in further view of Sheshappanavar et al. (Sheshappanavar et al., “PatchAugment: Local Neighborhood Augmentation in Point Cloud Classification”, 2021, 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), pages 2118-2127) (Hereinafter referred to as Sheshappanavar) in further view of Zhao et al. (Zhao et al., Divide and Conquer: 3D Point Cloud Instance Segmentation With Point-Wise Binarization, 20 Mar. 2023, pages 1-10) (Hereinafter referred to as Zhao).
Regarding claim 5, Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu fail to teach but Sheshappanavar teaches the method of claim 2, wherein the step (a) is further receiving scene data (Sheshappanavar; "We evaluate our method of patch augmentation using synthetic (ModelNet40 [10], ModelNet10 [10], and SHREC’16 [14]) and real-world (ScanObjectNN [11]) datasets," (page 2119, Introduction). "ScanObjectNN is a more complex dataset consisting of 2902 real-world objects spread across 15 classes. ScanObjectNN has 700 unique scenes developed from two popular scene object datasets i.e., SceneNN [44] and Scan-Net [8] with 100 and 1513 objects respectively… Each perturbed variant consists of five randomly sampled objects from the ground truth objects enlarging ScanObjectNN dataset into 14,510 perturbed objects," (Sheshappanavar; page 2123, 4.3 Datasets)), wherein the step (a-2) comprises sampling a plurality of anchor points targeting the scene data (Sheshappanavar; “This paper focuses on augmenting each sample’s different local neighborhoods... Our PatchAugment augments the neighborhood points as follows; first, it randomly drops a small fraction of the queried neighborhood points. Second, it randomly scales these neighborhood points, followed by a random perturb rotation by small angles and a random rotation along the up-axis direction. Then, it randomly translates each augmented neighborhood group. Note that this random translation is common to all points within a patch neighborhood but different for points belonging to different neighboring patches/grouped points,” (page 2119, 1. Introduction; page 2121 Fig. 2). Every point that does not get dropped is mapped to sampling a plurality of anchor points since the points in a neighborhood are used for local transformation. “We evaluate our method of patch augmentation using synthetic (ModelNet40 [10], ModelNet10 [10], and SHREC’16 [14]) and real-world (ScanObjectNN [11]) datasets,” (Sheshappanavar; page 2119, 1. Introduction). The patch augment method, which includes the aforementioned steps, is applied to ScanObjectNN data which contains scene data.), the step (a-2) further comprising: scene objects ((Sheshappanavar; page 2119, 1. Introduction; page 2121 Fig. 2). Every point that does not get dropped is mapped to sampling one or more anchor points since the points in a neighborhood are used for local transformation.
“In patch augmentation a sample object as shown in Figure 2(a) undergoes a series of DA techniques at patch level, i.e., random points dropout (λ), random scale (S), random rotations (both perturbed R and up-axis rotations Rθ), random translation (T), and random jittering (J) as shown in part (b) to (g) of Figure 2 respectively,” (Sheshappanavar; page 2120, 3. Method). The sample object reads on instance; thus, the patch augment method, including the steps that read on sampling one or more anchor points, is applied to an instance.
“From the several variants of ScanObjectNN dataset we have considered the six prominently used variants for our experiments… Each perturbed variant consists of five randomly sampled objects from the ground truth objects enlarging ScanObjectNN dataset into 14,510 perturbed objects,” (Sheshappanavar; page 2123, 4.3 Datasets). The patch augment method, including the steps that read on sampling one or more anchor points, is applied to ScanObjectNN data which contains a plurality of scene objects.).
Before the effective filling date of the claimed invention, it would have been obvious to one having ordinary skill in the art to apply the teachings of Sheshappanavar to Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu. The motivation would have been to “enhance the overall augmentation of the sample object, thereby improving the learning of local geometry by the deep neural network…,” (Sheshappanavar; page 2119, 1. Introduction).
Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu in further view of Sheshappanavar does not explicitly disclose segmenting a plurality of instances in the scene data, wherein the plurality of segmented instances correspond to the scene objects.
Zhao teaches segmenting a plurality of instances in the scene data, wherein the plurality of segmented instances correspond to the scene objects (Zhao; "…we propose a novel proposal generation framework to better segment adjacent objects and complete instances," (page 1, Introduction; page 3, Architecture Overview). Figure 6 shows more than one instance that has been segmented in a scene by Zhao's method (page 7, Fig. 6; page 5, Experiment Setting). Figure 6 shows that the segmented instances correspond to scene objects.).
Before the effective filling date of the claimed invention, it would have been obvious to one having ordinary skill in the art to apply the teachings of Zhao to Juppe in view of Kim in further view of Zhang in further view of Hacinecipoglu in further view of Sheshappanavar. The motivation would have been to allow for the application of object-based augmentations.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ERICA G THERKORN whose telephone number is (571)272-2939. The examiner can normally be reached Monday - Friday 9:00am - 5:00pm.
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, Devona Faulk can be reached at 571-272-7515. 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.
/ERICA G THERKORN/Examiner, Art Unit 2618
/DEVONA E FAULK/Supervisory Patent Examiner, Art Unit 2618