DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Priority
Receipt is acknowledged of certified copies of papers submitted under 35 U.S.C. 119(a)-(d), which papers have been placed of record in the file.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 11/13/2025 has been considered by the examiner.
Election/Restrictions
Applicant’s election without traverse of claims 1-14 in the reply filed on 04/16/2026 is acknowledged.
Drawing Objections
The drawings are objected to because:
In Figure 2, #150 called “preprocesso” should read “preprocessor” in order to correct spelling error.
In Figure 8, step S110 stating “extracting first global feature from D partial point coud of object” should read “extracting first global feature from 3D partial point cloud of object” in order to correct spelling errors.
Appropriate correction is required.
Claim Objections
Claim 8 is objected to because of the following informalities:
In claim 8, line 1, “A apparatus…” should read “An apparatus…” in order to use the proper indefinite article.
In claim 8, line 5, “when executed by the processor, cause the processor to:” should read “when executed by the processor, causes the processor to:” in order to correct the grammatical error.
Appropriate correction is required.
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of pre-AIA 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
(a)(2) the claimed invention was described in a patent issued under section 151, or in application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claims 1-6 and 8-13 are rejected under 35 U.S.C. 102(a)(1)/(a)(2) as being anticipated by FU et al. (US 20220292698 A1), hereinafter referenced as FU.
Regarding claim 1, FU teaches a method for estimating a pose of an object (Fig. 2, Paragraph [0040] – FU discloses a method for category-level 6D pose and size estimation.),
PNG
media_image1.png
529
1598
media_image1.png
Greyscale
Diagram of FU’s Fig. 2 illustrating the framework for category level 6D pose and size estimation.
the method comprising: obtaining a 3 dimensional (3D) partial point cloud including partial points among full points for the object (Fig. 1, Paragraph [0061] – FU discloses in this invention it is to estimate the 6D object pose and size of a set of unseen instances with known categories, presented by a partial point cloud. We represent the 6D object pose as a rigid-body homogeneous transformation matrix [R|t]∈SE(3), where rotation R∈SO(3) and translation t∈R3. SE(3) and SO(3) indicate the Lie group of 3D rigid transformations and 3D rotations, individually.)
from a 3D camera (Fig. 2, Paragraph [0099] – FU discloses we also deploy our model for inferring 6D object pose and size to execute manipulation tasks on a real Baxter robot which is a dual-arm collaborative robot with parallel grippers, mounted with a RealSense D435 Camera on the base [wherein RealSense D435 is a 3D camera].);
estimating a 3D full point cloud for the object from the 3D partial point cloud including the partial points (Fig. 2, Paragraph [0075] – FU discloses one remedy to rotate the partial point O(R, tn|c) by 180° around its symmetry-axis in the object frame, to generate the paired points O′(R, tn|c). It also enables our network to reason the occluded part from the observable one, to obtain a more complete shape for subsequent dimensional estimation. Paragraphs [0077-78] – FU further discloses we combine the observation O(R, tn|c) and ground truth symmetric points O′(R, tn|c) during training, namely the grouped points Gori(R, tn|c)=O(R, tn|c)+O′(R, tn|c). Such a combination step generates a relatively complete object shape [wherein the complete shape is a 3D full point cloud] for coarse center localization.)
using a pre-trained modeling converter (Fig. 2, Paragraph [0012] – FU discloses it is proposed a geometry-based approach for 6D object pose and size recovery, from a single depth image, without external pose-annotated real-world training data. Paragraph [0087] – FU further discloses our network is exclusively trained with synthetic depth images without any real pose-annotated images [wherein the network is the modeling converter].);
and estimating a 6 dimensional (6D) pose of the object from the estimated 3D full point cloud (Fig. 2, Paragraph [0064] – FU discloses FIG. 2 illustrates an overview of the network for category-level 6D pose and size estimation according to this invention. The pre-processing stage (left side) outputs the predicted category labels and potential masks of the target instances (mug as an example). The back-projected points from depth observation and the canonical category-specific keypoints are fed into the main network (right side). The network includes four output branches that generate the 3D-OCR, GeoReS, MPDV, and uniform scale.).
Regarding claim 2, FU teaches the method of claim 1,
PNG
media_image1.png
529
1598
media_image1.png
Greyscale
Diagram of FU’s Fig. 2 illustrating the framework for category level 6D pose and size estimation.
FU further teaches wherein the generating the 3D full point cloud of the object (Fig. 2, Paragraph [0080] – FU discloses compared to regressing only with partial points set O(R, tn|c), the combined points G(R, tn|c) provides a more complete shape [wherein a complete shape is a full point cloud] prior for more accurate size estimation as illustrated below.) includes:
extracting a first global feature of the object from the 3D partial point cloud using a first autoencoder included in the modeling converter (Fig. 2, Paragraph [0071] – FU discloses normalized partial observation [wherein partial observation is partial point cloud] O(R, tn|c) and the category-specific canonical template representation TK(R0, t0|c) are fed into the network and the encoders aim to extract feature FO and FT, respectively.);
transforming the first global feature to a second global feature using a latent space association network included in the modeling converter (Fig. 2, Paragraph [0071] – FU discloses for better aggregating low-level and high-level feature, our network conducts the maxpooling of each K-dimensional vector derived from the last n layers and concatenates them as a multiple latent feature FO=f1⊙f2 . . . ⊙fn, where ⊙ denotes concatenation.);
and generating the 3D full point cloud for the object by reconstructing the 3D partial point cloud based on the second global feature using a second autoencoder included in the modeling converter (Fig. 2, Paragraph [0028] – FU discloses the corresponding reflective observation implicitly decouples the pose from the original input depth observation by sending the second multiple latent feature into the second decoder for point-wise paired prediction. Paragraph [0029] – FU discloses in the MPDE module, a relative complete object shape is obtained by grouping the output of the GeoReS module and the original input depth observation.).
Regarding claim 3, FU teaches the method of claim 2,
FU further teaches wherein the first autoencoder is trained to receive the 3D partial point cloud including the partial points for the object (Fig. 2, Paragraph [0071] – FU discloses normalized partial observation O(R, tn|c) [wherein normalized partial observation is the 3D partial point cloud] and the category-specific canonical template representation TK(R0, t0|c) are fed into the network and the encoders aim to extract feature FO and FT, respectively.),
extract the first global feature of the object from the 3D partial point cloud including the partial points for the object (Fig. 2, Paragraph [0071] – FU discloses normalized partial observation O(R, tn|c) [wherein normalized partial observation is the 3D partial point cloud] and the category-specific canonical template representation TK(R0, t0|c) are fed into the network and the encoders aim to extract feature FO and FT, respectively.),
reconstruct the 3D partial point cloud including the partial points for the object based on the extracted first global feature (Fig. 2, Paragraph [0071] – FU discloses for better aggregating low-level and high-level feature, our network conducts the maxpooling of each K-dimensional vector derived from the last n layers and concatenates them as a multiple latent feature FO=f1⊙f2 . . . ⊙fn, where ⊙ denotes concatenation.),
and output the reconstructed 3D partial point cloud including the partial points for the object (Fig. 2, Paragraph [0071] – FU discloses FO and FT are concatenated by N repetition to generate per-point embedding before feeding into the decoder, which aims to perform orientation-guided characterization under perspective context extracted from the partial observations O(R,tn|c). See also Paragraph [0076].).
Regarding claim 4, FU teaches the method of claim 2,
FU further teaches wherein the second autoencoder is trained to receive the 3D full point cloud for the object (Fig. 2, Paragraph [0077] – FU discloses built upon the GeoReS module, we further obtain a relative complete object shape beneficial to center localization and size regression, by grouping the output O(
R
^
,
t
n
^
|
c
) of the GeoReS branch and input O(R, tn|c).),
extract the second global feature of the object from the 3D full point cloud (Fig. 2, Paragraph [0077] – FU discloses we combine the observation O(R, tn|c) and ground truth symmetric points O′(R, tn|c) during training, namely the grouped points Gori(R, tn|c)=O(R, tn|c)+O′(R, tn|c).),
reconstruct the 3D full point cloud based on the extracted first global feature (Fig. 2, Paragraph [0077] – FU discloses we combine the observation O(R, tn|c) and ground truth symmetric points O′(R, tn|c) during training, namely the grouped points Gori(R, tn|c)=O(R, tn|c)+O′(R, tn|c). Paragraph [0078] – FU discloses such a combination step generates a relatively complete object shape for coarse center localization.),
and output the reconstructed 3D full point cloud for the object (Fig. 2, Paragraph [0029] – FU discloses in the MPDE module, a relative complete object shape is obtained by grouping the output of the GeoReS module and the original input depth observation.).
Regarding claim 5, FU teaches the method of claim 2,
FU further teaches wherein the latent space association network is trained to receive the first global feature of the object from the first autoencoder (Fig. 2, Paragraph [0075] – FU discloses One remedy to rotate the partial point O(R, tn|c) by 180° around its symmetry-axis in the object frame, to generate the paired points O′(R, tn|c). It also enables our network to reason the occluded part from the observable one, to obtain a more complete shape for subsequent dimensional estimation.),
receive the second global feature of the object from the second autoencoder as label data (Fig. 2, Paragraph [0069] – FU discloses Given depth normalized observation O(R, tn|c), our goal is to standardize the shape-invariant rotation representation for each category-specific instance. That is to say, extract the representative orientation from the observation and map it onto the correspondent canonical template shape of the category c, inheriting the consistent orientation R.),
transform the first global feature to the second global feature (Fig. 2, Paragraph [0071] – FU discloses FO and FT are concatenated by N repetition to generate per-point embedding before feeding into the decoder, which aims to perform orientation-guided characterization under perspective context extracted from the partial observations O(R, tn|c).),
and output the second global feature transformed from the first global feature (Fig. 2, Paragraph [0077] – FU discloses we combine the observation O(R, tn|c) and ground truth symmetric points O′(R, tn|c) during training, namely the grouped points Gori(R, tn|c)=O(R, tn|c)+O′(R, tn|c). Paragraph [0078] – FU discloses such a combination step generates a relatively complete object shape for coarse center localization.).
Regarding claim 6, FU teaches the method of claim 1,
FU further teaches wherein the 3D full point cloud includes a 3D camera-based first coordinates (Fig. 2, Paragraph [0062] – FU discloses we denote the original input depth observation of a category-known instance as Oori(R, t|c)∈R3×N, where c is the category prior of the detected objects; N is the number of the valid back-projected points, and (R; t) is the pose described in the camera frame. In order to be robust against the global scale, for all input Oori(R, t|c) with different dimension, we shift and scale them to the unit sphere as O(R, tn|c) in Eq. 1.)
and an object-based second coordinates matched to the 3D camera-based first coordinates (Fig. 2, Paragraph [0069] – FU discloses given depth normalized observation O(R, tn|c), our goal is to standardize the shape-invariant rotation representation for each category-specific instance. That is to say, extract the representative orientation from the observation and map it onto the correspondent canonical template shape of the category c, inheriting the consistent orientation R.),
wherein the estimating the 6D pose of the object includes estimating the 6D pose of the object using a transformation matrix (Fig. 2, Paragraph [0061] – FU discloses in this invention it is to estimate the 6D object pose and size of a set of unseen instances with known categories, presented by a partial point cloud. We represent the 6D object pose as a rigid-body homogeneous transformation matrix [R|t]∈SE(3), where rotation R∈SO(3) and translation t∈R3. SE(3) and SO(3) indicate the Lie group of 3D rigid transformations and 3D rotations, individually. And the size of the object is formalized by scale factor s∈R.)
to minimize an error between the first coordinates and the second coordinates (Fig. 2, Paragraph [0013] – FU discloses the 3D-OCR (row b in FIG. 1) inherits the same orientation as the input partial observation (row a in FIG. 1), or we can assume that input partial observation and 3D-OCR are semantic aligned. See also Paragraph [0070].).
Regarding claim 8, FU teaches a apparatus for estimating a pose of an object (Fig. 2, Paragraph [0040] – FU discloses a method for category-level 6D pose and size estimation. Paragraph [0054] – FU further discloses embodiments of the subject matter described in this specification can, for example, be implemented as one or more computer program products, e.g., one or more modules of computer program instructions encoded on a computer readable medium for execution by, or to control the operation of, data processing apparatus. See also Paragraph [0055].), the apparatus comprising:
a memory configured to store a pose estimation program for estimating a pose of an object (Fig. 2, Paragraph [0055] – FU discloses computer readable medium can be a machine readable tangible storage device, a machine readable tangible storage substrate, a tangible memory device, or a combination of one or more of them. See also Paragraph [0057].);
and a processor configured to execute the pose estimation program stored in the memory (Fig. 2, Paragraph [0056] – FU discloses processes and logic flows described in this specification can be performed by one or more programmable processors executing one or more computer programs to perform functions by operating on input data and generating output. Paragraph [0057] – FU further discloses a processor will receive instructions and data from a read only memory or a random access memory or both.),
wherein the pose estimation program (Fig. 2, Paragraph [0056] – FU discloses processes and logic flows described in this specification can be performed by one or more programmable processors executing one or more computer programs to perform functions by operating on input data and generating output), when executed by the processor (Fig. 2, Paragraph [0056]),
cause the processor to: obtain a 3 dimensional (3D) partial point cloud including partial points among full points for the object (Fig. 1, Paragraph [0061] – FU discloses in this invention it is to estimate the 6D object pose and size of a set of unseen instances with known categories, presented by a partial point cloud. We represent the 6D object pose as a rigid-body homogeneous transformation matrix [R|t]∈SE(3), where rotation R∈SO(3) and translation t∈R3. SE(3) and SO(3) indicate the Lie group of 3D rigid transformations and 3D rotations, individually.)
from a 3D camera (Fig. 2, Paragraph [0099] – FU discloses we also deploy our model for inferring 6D object pose and size to execute manipulation tasks on a real Baxter robot which is a dual-arm collaborative robot with parallel grippers, mounted with a RealSense D435 Camera on the base [wherein RealSense D435 is a 3D camera].);
estimate a 6D full point cloud for the object from the 3D partial point cloud including the partial points (Fig. 2, Paragraph [0064] – FU discloses FIG. 2 illustrates an overview of the network for category-level 6D pose and size estimation according to this invention. The pre-processing stage (left side) outputs the predicted category labels and potential masks of the target instances (mug as an example). The back-projected points from depth observation and the canonical category-specific keypoints are fed into the main network (right side). The network includes four output branches that generate the 3D-OCR, GeoReS, MPDV, and uniform scale.)
using a pre-trained modeling converter (Fig. 2, Paragraph [0012] – FU discloses it is proposed a geometry-based approach for 6D object pose and size recovery, from a single depth image, without external pose-annotated real-world training data. Paragraph [0087] – FU further discloses our network is exclusively trained with synthetic depth images without any real pose-annotated images [wherein the network is the modeling converter].);
and estimate a 6 dimensional (6D) pose of the object from the estimated 6D full point cloud (Fig. 2, Paragraph [0064] – FU discloses FIG. 2 illustrates an overview of the network for category-level 6D pose and size estimation according to this invention. The pre-processing stage (left side) outputs the predicted category labels and potential masks of the target instances (mug as an example). The back-projected points from depth observation and the canonical category-specific keypoints are fed into the main network (right side). The network includes four output branches that generate the 3D-OCR, GeoReS, MPDV, and uniform scale.).
Regarding claim 9, FU teaches the apparatus of claim 8,
FU further teaches wherein the pre-trained modeling converter includes a first autoencoder and a second autoencoder connected with a latent space association network (Fig. 2, Paragraph [0071] – FU discloses our network uses PointNet-like structure as illustrated in FIG. 2. The normalized partial observation O(R, tn|c) and the category-specific canonical template representation TK(R0, t0|c) are fed into the network and the encoders aim to extract feature FO and FT, respectively. Specially, for better aggregating low-level and high-level feature, our network conducts the maxpooling of each K-dimensional vector derived from the last n layers and concatenates them as a multiple latent feature FO=f1⊙f2 . . . ⊙fn, where ⊙ denotes concatenation.),
and wherein the processor (Fig. 2, Paragraph [0056]) is configured to extract a first global feature of the object from the 3D partial point cloud using a first autoencoder included in the modeling converter (Fig. 2, Paragraph [0071] – FU discloses normalized partial observation [wherein partial observation is partial point cloud] O(R, tn|c) and the category-specific canonical template representation TK(R0, t0|c) are fed into the network and the encoders aim to extract feature FO and FT, respectively.),
transform the first global feature to a second global feature using the latent space association network included in the modeling converter (Fig. 2, Paragraph [0071] – FU discloses for better aggregating low-level and high-level feature, our network conducts the maxpooling of each K-dimensional vector derived from the last n layers and concatenates them as a multiple latent feature FO=f1⊙f2 . . . ⊙fn, where ⊙ denotes concatenation.),
generate the 3D full point cloud for the object by reconstructing the 3D partial point cloud based on the second global feature using a second autoencoder included in the modeling converter (Fig. 2, Paragraph [0028] – FU discloses the corresponding reflective observation implicitly decouples the pose from the original input depth observation by sending the second multiple latent feature into the second decoder for point-wise paired prediction. Paragraph [0029] – FU discloses in the MPDE module, a relative complete object shape is obtained by grouping the output of the GeoReS module and the original input depth observation.).
Regarding claim 10, FU teaches the apparatus of claim 9,
FU further teaches wherein the first autoencoder is trained to receive the 3D partial point cloud including the partial points for the object (Fig. 2, Paragraph [0071] – FU discloses normalized partial observation O(R, tn|c) [wherein normalized partial observation is the 3D partial point cloud] and the category-specific canonical template representation TK(R0, t0|c) are fed into the network and the encoders aim to extract feature FO and FT, respectively.),
extract the first global feature of the object from the 3D partial point cloud including the partial points for the object (Fig. 2, Paragraph [0071] – FU discloses normalized partial observation O(R, tn|c) [wherein normalized partial observation is the 3D partial point cloud] and the category-specific canonical template representation TK(R0, t0|c) are fed into the network and the encoders aim to extract feature FO and FT, respectively.),
reconstruct the 3D partial point cloud including the partial points for the object based on the extracted first global feature (Fig. 2, Paragraph [0071] – FU discloses for better aggregating low-level and high-level feature, our network conducts the maxpooling of each K-dimensional vector derived from the last n layers and concatenates them as a multiple latent feature FO=f1⊙f2 . . . ⊙fn, where ⊙ denotes concatenation.),
and output the reconstructed 3D partial point cloud including the partial points for the object (Fig. 2, Paragraph [0071] – FU discloses FO and FT are concatenated by N repetition to generate per-point embedding before feeding into the decoder, which aims to perform orientation-guided characterization under perspective context extracted from the partial observations O(R,tn|c). See also Paragraph [0076].).
Regarding claim 11, FU teaches the apparatus of claim 9,
FU further teaches wherein the second autoencoder is trained to receive the 3D full point cloud for the object (Fig. 2, Paragraph [0077] – FU discloses built upon the GeoReS module, we further obtain a relative complete object shape beneficial to center localization and size regression, by grouping the output O(
R
^
,
t
n
^
|
c
) of the GeoReS branch and input O(R, tn|c).),
extract the second global feature of the object from the 3D full point cloud (Fig. 2, Paragraph [0077] – FU discloses we combine the observation O(R, tn|c) and ground truth symmetric points O′(R, tn|c) during training, namely the grouped points Gori(R, tn|c)=O(R, tn|c)+O′(R, tn|c).),
reconstruct the 3D full point cloud based on the extracted first global feature (Fig. 2, Paragraph [0077] – FU discloses we combine the observation O(R, tn|c) and ground truth symmetric points O′(R, tn|c) during training, namely the grouped points Gori(R, tn|c)=O(R, tn|c)+O′(R, tn|c). Paragraph [0078] – FU discloses such a combination step generates a relatively complete object shape for coarse center localization.),
and output the reconstructed 3D full point cloud for the object (Fig. 2, Paragraph [0029] – FU discloses in the MPDE module, a relative complete object shape is obtained by grouping the output of the GeoReS module and the original input depth observation.).
Regarding claim 12, FU teaches the apparatus of claim 9,
FU further teaches wherein the latent space association network is trained to receive the first global feature of the object from the first autoencoder (Fig. 2, Paragraph [0075] – FU discloses One remedy to rotate the partial point O(R, tn|c) by 180° around its symmetry-axis in the object frame, to generate the paired points O′(R, tn|c). It also enables our network to reason the occluded part from the observable one, to obtain a more complete shape for subsequent dimensional estimation.),
receive the second global feature of the object from the second autoencoder as label data (Fig. 2, Paragraph [0069] – FU discloses Given depth normalized observation O(R, tn|c), our goal is to standardize the shape-invariant rotation representation for each category-specific instance. That is to say, extract the representative orientation from the observation and map it onto the correspondent canonical template shape of the category c, inheriting the consistent orientation R.),
transform the first global feature to the second global feature (Fig. 2, Paragraph [0071] – FU discloses FO and FT are concatenated by N repetition to generate per-point embedding before feeding into the decoder, which aims to perform orientation-guided characterization under perspective context extracted from the partial observations O(R, tn|c).),
and output the second global feature transformed from the first global feature (Fig. 2, Paragraph [0077] – FU discloses we combine the observation O(R, tn|c) and ground truth symmetric points O′(R, tn|c) during training, namely the grouped points Gori(R, tn|c)=O(R, tn|c)+O′(R, tn|c). Paragraph [0078] – FU discloses such a combination step generates a relatively complete object shape for coarse center localization.).
Regarding claim 13, FU teaches the apparatus of claim 9,
FU further teaches wherein the 3D full point cloud includes a 3D camera-based first coordinates (Fig. 2, Paragraph [0062] – FU discloses we denote the original input depth observation of a category-known instance as Oori(R, t|c)∈R3×N, where c is the category prior of the detected objects; N is the number of the valid back-projected points, and (R; t) is the pose described in the camera frame. In order to be robust against the global scale, for all input Oori(R, t|c) with different dimension, we shift and scale them to the unit sphere as O(R, tn|c) in Eq. 1.)
and an object-based second coordinates matched to the 3D camera-based first coordinates (Fig. 2, Paragraph [0069] – FU discloses given depth normalized observation O(R, tn|c), our goal is to standardize the shape-invariant rotation representation for each category-specific instance. That is to say, extract the representative orientation from the observation and map it onto the correspondent canonical template shape of the category c, inheriting the consistent orientation R.),
wherein the processor is configured to estimate the 6D pose of the object using a transformation matrix (Fig. 2, Paragraph [0061] – FU discloses in this invention it is to estimate the 6D object pose and size of a set of unseen instances with known categories, presented by a partial point cloud. We represent the 6D object pose as a rigid-body homogeneous transformation matrix [R|t]∈SE(3), where rotation R∈SO(3) and translation t∈R3. SE(3) and SO(3) indicate the Lie group of 3D rigid transformations and 3D rotations, individually. And the size of the object is formalized by scale factor s∈R.)
to minimize an error between the first coordinates and the second coordinates (Fig. 2, Paragraph [0013] – FU discloses the 3D-OCR (row b in FIG. 1) inherits the same orientation as the input partial observation (row a in FIG. 1), or we can assume that input partial observation and 3D-OCR are semantic aligned. See also Paragraph [0070].).
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed
invention is not identically disclosed as set forth in section 102 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 7 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over FU (US 20220292698 A1), hereinafter referenced as FU in view of ORMAN (US 5767960 A), hereinafter referenced as ORMAN.
Regarding claim 7, FU teaches the method of claim 6,
FU fails to explicitly teach wherein the 6D pose includes rotation angles around three directions of the object and translation distances along the three directions of the object.
However, ORMAN explicitly teaches wherein the 6D pose (Fig. 1, Col. 1, Lines [13-14] – ORMAN discloses the device is designed for measuring position and orientation in six degrees of freedom.)
includes rotation angles around three directions of the object (Fig. 1, Col. 1, Lines [13-22] – ORMAN discloses the device is designed for measuring position and orientation in six degrees of freedom, namely motion or translation in three coordinate directions (location), and rotational motion about three coordinate axes (orientation), location being commonly defined as a set of x, y, and z linear coordinates referring to three mutually perpendicular directions (axes) and orientation being commonly defined as pitch, roll and azimuth angular coordinates about three mutually perpendicular axes usually coincident with three mutually perpendicular directions.)
and translation distances along the three directions of the object (Fig. 1, Col. 1, Lines [13-22] – ORMAN discloses the device is designed for measuring position and orientation in six degrees of freedom, namely motion or translation in three coordinate directions (location), and rotational motion about three coordinate axes (orientation), location being commonly defined as a set of x, y, and z linear coordinates referring to three mutually perpendicular directions (axes) and orientation being commonly defined as pitch, roll and azimuth angular coordinates about three mutually perpendicular axes usually coincident with three mutually perpendicular directions.).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date the claimed invention was made to combine the teachings of FU of having a method for estimating a pose of an object, the method comprising: obtaining a 3 dimensional (3D) partial point cloud including partial points among full points for the object from a 3D camera; estimating a 3D full point cloud for the object from the 3D partial point cloud including the partial points using a pre-trained modeling converter; and estimating a 6 dimensional (6D) pose of the object from the estimated 3D full point cloud, with the teachings of ORMAN having wherein the 6D pose includes rotation angles around three directions of the object and translation distances along the three directions of the object.
Wherein FU’s method wherein the 6D pose includes rotation angles around three directions of the object and translation distances along the three directions of the object.
The motivation behind this modification would have been to provide an enhanced pose estimation method with improved accuracy and reduced positional errors, since both FU and ORMAN relate to spatial parameter determination of objections, wherein FU relates to a network of category-level 6D object pose and size estimation; a new MPDE component is proposed for accurate object center and size estimation, and jointly training modules aforementioned further improves the performance, and ORMAN relates to the determination of the six spatial parameters position (x, y, z) and orientation (ψ, θ, φ) of an object relative to a fixed reference point; this invention employs techniques to ensure transmitter scanning stability, as well as unique timing measurement techniques to achieve accurate spatial coordinate measurements.. Please see FU (US 20220292698 A1), Paragraph [0044], and ORMAN (US 5767960 A), Col. 3, Lines [6-12].
Regarding claim 14, FU teaches the apparatus of claim 13,
FU fails to explicitly teach wherein the 6D pose includes rotation angles around three directions of the object and translation distances along the three directions of the object.
However, ORMAN explicitly teaches wherein the 6D pose (Fig. 1, Col. 1, Lines [13-14] – ORMAN discloses the device is designed for measuring position and orientation in six degrees of freedom.)
includes rotation angles around three directions of the object (Fig. 1, Col. 1, Lines [13-22] – ORMAN discloses the device is designed for measuring position and orientation in six degrees of freedom, namely motion or translation in three coordinate directions (location), and rotational motion about three coordinate axes (orientation), location being commonly defined as a set of x, y, and z linear coordinates referring to three mutually perpendicular directions (axes) and orientation being commonly defined as pitch, roll and azimuth angular coordinates about three mutually perpendicular axes usually coincident with three mutually perpendicular directions.)
and translation distances along the three directions of the object (Fig. 1, Col. 1, Lines [13-22] – ORMAN discloses the device is designed for measuring position and orientation in six degrees of freedom, namely motion or translation in three coordinate directions (location), and rotational motion about three coordinate axes (orientation), location being commonly defined as a set of x, y, and z linear coordinates referring to three mutually perpendicular directions (axes) and orientation being commonly defined as pitch, roll and azimuth angular coordinates about three mutually perpendicular axes usually coincident with three mutually perpendicular directions.).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date the claimed invention was made to combine the teachings of FU of having a apparatus for estimating a pose of an object, the apparatus comprising: a memory configured to store a pose estimation program for estimating a pose of an object; and a processor configured to execute the pose estimation program stored in the memory, wherein the pose estimation program, when executed by the processor, cause the processor to: obtain a 3 dimensional (3D) partial point cloud including partial points among full points for the object from a 3D camera; estimate a 6D full point cloud for the object from the 3D partial point cloud including the partial points using a pre-trained modeling converter; and estimate a 6 dimensional (6D) pose of the object from the estimated 6D full point cloud, with the teachings of ORMAN having wherein the 6D pose includes rotation angles around three directions of the object and translation distances along the three directions of the object.
Wherein FU’s apparatus wherein the 6D pose includes rotation angles around three directions of the object and translation distances along the three directions of the object.
The motivation behind this modification would have been to provide an enhanced pose estimation apparatus with improved accuracy and reduced positional errors, since both FU and ORMAN relate to spatial parameter determination of objections, wherein FU relates to a network of category-level 6D object pose and size estimation; a new MPDE component is proposed for accurate object center and size estimation, and jointly training modules aforementioned further improves the performance, and ORMAN relates to the determination of the six spatial parameters position (x, y, z) and orientation (ψ, θ, φ) of an object relative to a fixed reference point; this invention employs techniques to ensure transmitter scanning stability, as well as unique timing measurement techniques to achieve accurate spatial coordinate measurements.. Please see FU (US 20220292698 A1), Paragraph [0044], and ORMAN (US 5767960 A), Col. 3, Lines [6-12].
Conclusion
Listed below are the prior arts made of record and not relied upon but are considered pertinent to applicant’s disclosure.
TREMBLAY et al. (US 20220379484 A1) - Apparatuses, systems, and techniques generate poses of an object based on data of the object observed from a first viewpoint and a second viewpoint. The poses can be evaluated to determine a portion of the data usable by an estimator to generate a pose of the object......… Fig. 1, Abstract.
CHANDLER et al. (US 20220277515 A1) - A computer-implemented method of modelling a common structure component, the method comprising, in a modelling computer system: receiving a plurality of captured frames, each frame comprising a set of 3D structure points, in which at least a portion of a common structure component is captured; computing a first reference position within at least one first frame of the plurality of frames; selectively extracting first 3D structure points of the first frame based on the first reference position computed for the first frame; computing a second reference position within a second frame of the plurality of frames; selectively extracting second 3D structure points of the second frame based on the second reference position computed for the second frame; and aggregating the first 3D structure points and the second 3D structure points, thereby generating an aggregate 3D model of the common structure component based on the first and second reference positions.........… Fig. 1, Abstract.
KAYSER et al. (US 20230124868 A1) - A computer-implemented method for machine vision includes generating, by an artificial neural network, a 6D position estimate of an object and a shape completion estimate of the object based on a partial point cloud. The partial point cloud including a set of 3D points representing the object. The artificial neural network trained using at least one digital geometry model of the object, in particular, at least one CAD model of the object.....… Fig. 1, Abstract.
LI et al. (US 20220164565 A1) - A method with object detection includes: obtaining a first point cloud feature based on point cloud data of an image; and determining at least one object in the image based on the first point cloud feature.........… Fig. 1, Abstract.
MOUSAVIAN et al. (US 20200363815 A1) - A method with object detection includes: obtaining a first point cloud feature based on point cloud data of an image; and determining at least one object in the image based on the first point cloud feature.........… Fig. 1, Abstract.
Lin, Haitao, et al. "Sar-net: Shape alignment and recovery network for category-level 6d object pose and size estimation." Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2022.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to BEZAWIT N SHIMELES whose telephone number is (571)272-7663. The examiner can normally be reached M-F 7:30am-5pm.
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, Chineyere Wills-Burns can be reached at (571) 272-9752. 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.
/BEZAWIT NOLAWI SHIMELES/Examiner, Art Unit 2673
/CHINEYERE WILLS-BURNS/Supervisory Patent Examiner, Art Unit 2673
Prosecution Insights