Notice of Pre-AIA or AIA Status
1. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
2. The Office Action is in response to amendment filed on 12/02/2025.
Information Disclosure Statement
3. The information disclosure statements (IDS) submitted on 09/16/2025 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statements are being considered by the examiner.
Response to Amendment
4. The amendment filed on 12/02/2025, independent claims 1 and 13 have been amended to incorporate subject matter from original claim 7, amends claims 3, 6, 9, 11, 15, and 16, and adds new claims 21-24; therefore, Claims 1-6, 8-16, 21-24 are pending.
5. Response to Arguments
Applicant’s arguments filed on 12/02/2025, pages 8-15 have been fully considered.
Claim Rejections - 35 USC §112
Since applicant has amended the claims properly, the 112(b) rejection in the non-final office action 09/12/2025 is withdrawn.
Claim Rejections - 35 USC §102
Since application has cancelled independent claim 17 and its depend claims 18-20, the 102 rejection in the non-final office action 09/12/2025 is moot.
Claim Rejections - 35 USC §102
Since application has incorporated limitations in previous claim 7 into independent claims, the 102 rejection on independent claims in the non-final office action 09/12/2025 is withdrawn.
Claim Rejections - 35 USC §103
Applicant’s arguments with respect to claim under 35 U.S.C. § 103 has been fully considered.
Applicant argued that the current prior arts (YOO et al. (WO 2024101635) and in view of DU (US 20190087686)) does not discloses the amended limitations as: “transforming the first type of facial feature data into a second type of facial feature data…. wherein the first type of facial feature data and the second type of facial feature data are organized in two different formats selected from the following: 2D Landmarks, 2D keypoints, 3D keypoints, segmentation map, compact feature, or facial semantics”, since:
“Yoo discloses "[t]he face swapping framework of the present invention is a lightweight
one-stage framework that can generate face-swapped images in real time without additional networks or processing through a novel decoder structure, data amplification, and switch-test strategy." See Yoo at para. [0071] (emphases added). Yoo also discloses "[w]hen the face swapping framework 200 receives a source image (Xs) and a target image (Xt) and generates a face-transformed image (Y), it additionally receives an attribute image (Xatt). The attribute image(Xatt) is an image for extracting the attributes of the target image (Xt), and may be an image obtained by adjusting the size of the source image (Xs) or the target image (Xt)." See id. at para.[0073] (emphases added). Yoo also discloses "[t]he adaptive normalization block 231 performs three adaptive normalizations considering the dimensions of each feature to guide the combination of identity and pose. Since the adaptive normalization block 231 performs adaptive normalization three times, it can be called a "Triple Adaptive Normalization (TAN) block." See id. at para. [0080] (emphases added). Yoo further discloses "[t]he spatially-adaptive pose integration block 231a includes a 1x1 convolutional layer, a pose activation function (P) and a ReLU activation function (ReLU). The pose activation function (P) is the kth pose feature (2*1,pose), normalized… According to Yoo, Yoo discloses a face swapping framework using an adaptive normalization block 231 to perform three adaptive normalizations considering the dimensions of each feature to guide the combination of identity and pose based on a source image (Xs), a target image (Xt), and an attribute image (Xatt). Yoo's normalization is performed by Spatially pose integration 231a, non-spatially pose integration 231b, and identity integration 231c on h in k , an input value of the… kth adaptive normalization block 231. However, Yoo does not disclose transforming the to another type of facial feature data”;
“ let alone "the first type of facial feature data and the second type of facial feature data are organized in two different formats selected from the following: 2D Landmarks, 2D keypoints, 3D keypoints, segmentation map, compact feature, or facial semantics." as recited in amended claim 1”;
“Du does not appear to contemplate "transforming the first type of facial feature data into a
second type of facial feature data." Therefore, Du fails to cure the deficiencies of Yoo”
Examiner’s Response:
The current claim languages in independent claims does not clarify what kind of transformation is performed and does not give any restriction of the transformation. Face swap can be interpreted as one kinds of face feature transformation. Yoo discloses "transforming the first type of facial feature data into a second type of facial feature data." In fig. 3/fig. 4, for example in which, the 1x1 conv. performs some transformation, which can interpreted as that the first type of facial feature (the input to the 1x1 Conv.) is transformed into the second type of facial feature (the output from the 1x1 Conv.); page 10, The spatially-adaptive pose integration block 231a includes a 1x1 convolutional layer, a pose activation function (P) and a ReLU activation function (ReLU). The pose activation function (P) is the kth pose feature ( ) normalized using 2D adaptive parameters generated from Denormalize . and the pose activation function (P) is as shown in [Equation 1];
Yoo also discloses the limitations in fig. 1 and fig. 2; for example, in fig. 2, the input to TAN 230 can be interpreted as the first type of facial feature; and the output from TAN 230 can be interpreted as the second type of facial feature and 230 perform the transformation.
Therefore, the combination of YOO and DU teaches the amended limitations of: “: “transforming the first type of facial feature data into a second type of facial feature data…. wherein the first type of facial feature data and the second type of facial feature data are organized in two different formats selected from the following: 2D Landmarks, 2D keypoints, 3D keypoints, segmentation map, compact feature, or facial semantics” in independent claims.
Claim Rejections - 35 USC § 103
6. 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.
7. Claims 1, 13, 21 are rejected are rejected under 35 U.S.C. 103 as being unpatentable over YOO et al. (WO 2024101635) and in view of DU (US 20190087686).
.
Regarding claim 1, YOO teaches a method of decoding a bitstream to output one or more pictures for a video stream (fig. 2; page 5, …. and a TAN (TAN) that generates a face-transformed image by integrating the identity features, target code, and pose features. Includes Triple Adaptive Normalization decoder), the method comprising:
receiving a bitstream (fig. 2, decoder 230 receives bitstream from encoder 210) associated with a first type of facial feature data representing a facial picture (fig. 2, Xatt);
and decoding, using coded information of the bitstream (fig. 4, Zkt, pose is the coded information), one or more pictures (fig. 2, Y is one or more pictures; page 7, … and a target code ( ) and encode it with the target code ( ) by decoding the pose features ( , ) and a pose network (220) that extracts identity features ( , )), wherein the decoding comprises:
transforming the first type of facial feature data into a second type of facial feature data (as shown in fig. 3/fig. 4, the 1x1 conv. performs some transformation, which can interpreted as that the first type of facial feature (the input to the 1x1 Conv.) is transformed into the second type of facial feature (the output from the 1x1 Conv.); page 10, The spatially-adaptive pose integration block 231a includes a 1x1 convolutional layer, a pose activation function (P) and a ReLU activation function (ReLU). The pose activation function (P) is the kth pose feature ( ) normalized using 2D adaptive parameters generated from Denormalize . and the pose activation function (P) is as shown in [Equation 1];
also in fig. 1 and fig. 2; for example, in fig. 2, the input to TAN 230 can be interpreted as the first type of facial feature; and the output from TAN 230 can be interpreted as the second type of facial feature and 230 perform the transformation.
);
and reconstructing the facial picture based on the second type of facial feature data (as shown in fig. 2, image Y; page 13, … target code reconstructed by inputting the face-transformed image (Y) into the encoder of the pose network 220. Since the face-transformed image (Y) has the same pose as the target image (X .sub.t )).
It is noticed that YOO does not disclose explicitly of wherein the first type of facial feature data and the second type of facial feature data are organized in two different formats selected from the following: 2D Landmarks, 2D keypoints, 3D keypoints, segmentation map, compact feature, or facial semantics.
DU discloses of wherein the first type of facial feature data and the second type of facial feature data are organized in two different formats selected from the following: 2D Landmarks, 2D keypoints, 3D keypoints, segmentation map, compact feature, or facial semantics (paragraph 0064, …multiply the facial feature information by the semantic feature information based on corresponding elements to generate the combined feature information; paragraph 0006, …Each point in the first facial feature map is used to represent a confidence level of the human face located in a region of the to-be-detected image corresponding to the each point of the first facial feature map. Each point in each of the plurality of second facial feature maps is used to represent position information of a region of the to-be-detected image corresponding to the each point of the second facial feature map. The first facial feature map and the second facial feature maps are respectively represented by matrixs; which is the key point.).
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to incorporate the technology that wherein the first type of facial feature data and the second type of facial feature data are organized in two different formats selected from the following: 2D Landmarks, 2D keypoints, 3D keypoints, segmentation map, compact feature, or facial semantics as a modification to the method for the benefit of that to get the facial data more accuracy and effectively (paragraph 0064).
Regarding claim 13, YOO teaches a method of encoding a video sequence into a bitstream (fig. 2; page 7, …. Identity encoder 210 generates first and second identity features (X s ) from the source image (X .sub.s ). , )), the method comprising:
receiving a video sequence (fig. 2, Xs is video sequence);
encoding one or more pictures of the video sequence (fig. 2, encoder 210 encode videos);
and generating a bitstream associated with the encoded pictures (fig. 2, the bitstream after encoder 210), wherein the encoding comprises:
generating a first type of facial feature data representing a facial picture (page 7-8, …Identity encoder 210 generates first and second identity features (X s ) from the source image (X .sub.s ). , ) and the first identity feature ( ) from the identity feature input value ( ) includes a convolutional layer that extracts. More specifically, the identity encoder 210 uses two down-sampling blocks to generate .sub.a first identity feature ( ) is extracted, and the second identity feature ( ) is extracted);
and generating and encoding, into the bitstream, information associated with the first type of facial feature data (as shown in fig. 2, Xatt; page 8, … Identity encoder 210 encodes the first identity feature ( ) as the first identity feature ( ), the identity feature input value ( ) to obtain. First and second identity characteristics ( , ) and identity feature input values ( ) is input to the TAN decoder 230) , wherein the first type of facial feature data is transformable by a translator into a second type of facial feature data (as shown in fig. 3/fig. 4, the 1x1 conv. performs some transformation, which can interpreted as that the first type of facial feature (the input to the 1x1 Conv.) is transformed into the second type of facial feature (the output from the 1x1 Conv.); page 10, The spatially-adaptive pose integration block 231a includes a 1x1 convolutional layer, a pose activation function (P) and a ReLU activation function (ReLU). The pose activation function (P) is the kth pose feature ( ) normalized using 2D adaptive parameters generated from Denormalize . and the pose activation function (P) is as shown in [Equation 1];
also in fig. 1 and fig. 2; for example, in fig. 2, the input to TAN 230 can be interpreted as the first type of facial feature; and the output from TAN 230 can be interpreted as the second type of facial feature and 230 perform the transformation.);
It is noticed that YOO does not disclose explicitly of wherein the first type of facial feature data and the second type of facial feature data are organized in two different formats selected from the following: 2D Landmarks, 2D keypoints, 3D keypoints, segmentation map, compact feature, or facial semantics.
DU discloses of wherein the first type of facial feature data and the second type of facial feature data are organized in two different formats selected from the following: 2D Landmarks, 2D keypoints, 3D keypoints, segmentation map, compact feature, or facial semantics (paragraph 0064, …multiply the facial feature information by the semantic feature information based on corresponding elements to generate the combined feature information; paragraph 0006, …Each point in the first facial feature map is used to represent a confidence level of the human face located in a region of the to-be-detected image corresponding to the each point of the first facial feature map. Each point in each of the plurality of second facial feature maps is used to represent position information of a region of the to-be-detected image corresponding to the each point of the second facial feature map. The first facial feature map and the second facial feature maps are respectively represented by matrixs; which is the key point.).
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to incorporate the technology that wherein the first type of facial feature data and the second type of facial feature data are organized in two different formats selected from the following: 2D Landmarks, 2D keypoints, 3D keypoints, segmentation map, compact feature, or facial semantics as a modification to the method for the benefit of that to get the facial data more accuracy and effectively (paragraph 0064).
Regarding claim 21, YOO teaches a method of signaling a bitstream (fig. 2; page 7, …. Identity encoder 210 generates first and second identity features (X s ) from the source image (X .sub.s ). , )), the method comprising:
receiving a video sequence (fig. 2, Xs is video sequence);
encoding one or more pictures of the video sequence (fig. 2, encoder 210 encode videos) by:
generating a first type of facial feature data representing a facial picture (page 7-8, …Identity encoder 210 generates first and second identity features (X s ) from the source image (X .sub.s ). , ) and the first identity feature ( ) from the identity feature input value ( ) includes a convolutional layer that extracts. More specifically, the identity encoder 210 uses two down-sampling blocks to generate .sub.a first identity feature ( ) is extracted, and the second identity feature ( ) is extracted);
and generating and encoding, into the bitstream, information associated with the first type of facial feature data (as shown in fig. 2, Xatt; page 8, … Identity encoder 210 encodes the first identity feature ( ) as the first identity feature ( ), the identity feature input value ( ) to obtain. First and second identity characteristics ( , ) and identity feature input values ( ) is input to the TAN decoder 230) , wherein the first type of facial feature data is transformable by a translator into a second type of facial feature data (as shown in fig. 3/fig. 4, the 1x1 conv. performs some transformation, which can interpreted as that the first type of facial feature (the input to the 1x1 Conv.) is transformed into the second type of facial feature (the output from the 1x1 Conv.); page 10, The spatially-adaptive pose integration block 231a includes a 1x1 convolutional layer, a pose activation function (P) and a ReLU activation function (ReLU). The pose activation function (P) is the kth pose feature ( ) normalized using 2D adaptive parameters generated from Denormalize . and the pose activation function (P) is as shown in [Equation 1];
also in fig. 1 and fig. 2; for example, in fig. 2, the input to TAN 230 can be interpreted as the first type of facial feature; and the output from TAN 230 can be interpreted as the second type of facial feature and 230 perform the transformation.);
and signaling a bitstream that is generated based on the encoding (FIG. 2, Y; also in fig. 12).
It is noticed that YOO does not disclose explicitly of wherein the first type of facial feature data and the second type of facial feature data are organized in two different formats selected from the following: 2D Landmarks, 2D keypoints, 3D keypoints, segmentation map, compact feature, or facial semantics.
DU discloses of wherein the first type of facial feature data and the second type of facial feature data are organized in two different formats selected from the following: 2D Landmarks, 2D keypoints, 3D keypoints, segmentation map, compact feature, or facial semantics (paragraph 0064, …multiply the facial feature information by the semantic feature information based on corresponding elements to generate the combined feature information; paragraph 0006, …Each point in the first facial feature map is used to represent a confidence level of the human face located in a region of the to-be-detected image corresponding to the each point of the first facial feature map. Each point in each of the plurality of second facial feature maps is used to represent position information of a region of the to-be-detected image corresponding to the each point of the second facial feature map. The first facial feature map and the second facial feature maps are respectively represented by matrixs; which is the key point.).
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to incorporate the technology that wherein the first type of facial feature data and the second type of facial feature data are organized in two different formats selected from the following: 2D Landmarks, 2D keypoints, 3D keypoints, segmentation map, compact feature, or facial semantics as a modification to the method for the benefit of that to get the facial data more accuracy and effectively (paragraph 0064).
8. Claim 2 is rejected are rejected under 35 U.S.C. 103 as being unpatentable over YOO et al. (WO 2024101635) and in view of DU (US 20190087686) and further in view of KANG et al. (WO 2022260386).
Regarding claim 2, the combination of YOO and DU teaches the limitations recited in claim 1 as discussed above.
It is noticed that YOO does not disclose explicitly of pre-processing the first type of facial feature data to generate a first feature vector that applies to a translator; translating, by the translator, the first feature vector into a second feature vector being associated with the second type of facial feature data; and post-processing the second feature vector to generate the second type of facial feature data.
KANG discloses of pre-processing the first type of facial feature data (i.e., process original facial feature vector mask) to generate a first feature vector (i.e., composite mask) that applies to a translator (page 14, … generating a composite mask from the original facial feature vector mask and the converted facial feature vector mask);
translating, by the translator, the first feature vector into a second feature vector (i.e., The second original facial feature vector) being associated with the second type of facial feature data (page 14, The second original facial feature vector is extracted by a convolutional product of the first original facial feature vector and the composite mask); and post-processing (i.e. convolution) the second feature vector to generate the second type of facial feature data (fig. 1, page 14, … and the second converted facial feature vector is extracted by a convolutional product of the first transformed facial feature vector and the composite mask).
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to incorporate the technology that pre-processing the first type of facial feature data to generate a first feature vector that applies to a translator; translating, by the translator, the first feature vector into a second feature vector being associated with the second type of facial feature data; and post-processing the second feature vector to generate the second type of facial feature data as a modification to the method for the benefit of that for lightening a deep learning model of an end-to-end deep learning network synthesizing a background and a face (page 3).
9. Claim 8, 14, 22 are rejected are rejected under 35 U.S.C. 103 as being unpatentable over YOO et al. (WO 2024101635) and in view of DU (US 20190087686) and further in view of CHRONQVIST (US 20100214445).
Regarding claim 8, the combination of YOO and DU teaches the limitations recited in claim 1 as discussed above.
It is noticed that YOO does not disclose explicitly of the bitstream comprises entropy coded data of the first type of facial feature data, and the decoding further comprises: entropy decoding the bitstream to obtain the first type of facial feature data.
CHRONQVIST discloses of the bitstream comprises entropy coded data of the first type of facial feature data, and the decoding further comprises: entropy decoding the bitstream to obtain the first type of facial feature data (paragraph 0024, … The image data can be encoded before storing with an image compression algorithm, e.g. a lossless compression algorithm such as run-length encoding, differential pulse-code modulation, predictive coding or entropy coding).
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to incorporate the technology that the bitstream comprises entropy coded data of the first type of facial feature data, and the decoding further comprises: entropy decoding the bitstream to obtain the first type of facial feature data as a modification to the method for the benefit of that for coding effectively (paragraph 0024).
Regarding claim 14, the combination of YOO and DU teaches the limitations recited in claim 13 as discussed above.
It is noticed that YOO does not disclose explicitly of the bitstream comprises entropy coded data of the first type of facial feature data.
CHRONQVIST discloses of the bitstream comprises entropy coded data of the first type of facial feature data (paragraph 0024, … The image data can be encoded before storing with an image compression algorithm, e.g. a lossless compression algorithm such as run-length encoding, differential pulse-code modulation, predictive coding or entropy coding).
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to incorporate the technology that the bitstream comprises entropy coded data of the first type of facial feature data, and the decoding further comprises: entropy decoding the bitstream to obtain the first type of facial feature data as a modification to the method for the benefit of that for coding effectively (paragraph 0024).
Regarding claim 22, the combination of YOO and DU teaches the limitations recited in claim 21 as discussed above.
It is noticed that YOO does not disclose explicitly of the bitstream comprises entropy coded data of the first type of facial feature data.
CHRONQVIST discloses of the bitstream comprises entropy coded data of the first type of facial feature data (paragraph 0024, … The image data can be encoded before storing with an image compression algorithm, e.g. a lossless compression algorithm such as run-length encoding, differential pulse-code modulation, predictive coding or entropy coding).
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to incorporate the technology that the bitstream comprises entropy coded data of the first type of facial feature data, and the decoding further comprises: entropy decoding the bitstream to obtain the first type of facial feature data as a modification to the method for the benefit of that for coding effectively (paragraph 0024).
10. Claim 3, 15, 16, 23, 24 are rejected are rejected under 35 U.S.C. 103 as being unpatentable over YOO et al. (WO 2024101635) and in view of DU (US 20190087686) and further in view of HE et al. (CN 115830414).
Regarding claim 3, the combination of YOO and DU teaches the limitations recited in claim 1 as discussed above.
It is noticed that YOO does not disclose explicitly of flattening all elements of the first type of facial feature data to obtain flattened data; and concatenating the flattened data to generate the first feature vector.
HE discloses of flattening all elements of the first type of facial feature data to obtain flattened data(fig. 2, step S203; page 7, … the flattening layer pulls the human face feature drawing output by the main module as the one-dimensional convolution feature with length k);
and concatenating the flattened data to generate the first feature vector (fig. 2, S203; page 7, …the neck module comprises a flattening layer and a full connecting layer, wherein the flattening layer pulls the human face feature drawing output by the main module as the one-dimensional convolution feature with length k, For example, the three-dimensional or four-dimensional human face feature matrix is pulled into a one-dimensional feature vector (i.e., a convolution feature vector); then, using the full connection layer to transform the convolution characteristic into one-dimensional representation characteristic, namely using the full connection layer to transform the one-dimensional characteristic vector after leveling; in which, the full connecting layer concatenating the flattened data to generate the first feature vector). It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to incorporate the technology that flattening all elements of the first type of facial feature data to obtain flattened data; and concatenating the flattened data to generate the first feature vector as a modification to the method for the benefit of that for getting some useful characteristics (page 7).
Regarding claim 15, the combination of YOO and DU teaches the limitations recited in claim 13 as discussed above.
It is noticed that YOO does not disclose explicitly of the bitstream comprises entropy coded data of flattened data that is generated by flattening all elements of the first type of facial feature data.
HE discloses of the entropy coded data of flattened data that is generated by flattening all elements of the first type of facial feature data (fig. 2, step S203; page 7, … the flattening layer pulls the human face feature drawing output by the main module as the one-dimensional convolution feature with length k).
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to incorporate the technology that entropy coded data of flattened data that is generated by flattening all elements of the first type of facial feature data as a modification to the method for the benefit of that for getting some useful characteristics (page 7).
Regarding claim 16, the combination of YOO and DU teaches the limitations recited in claim 13 as discussed above.
It is noticed that YOO does not disclose explicitly of flattening all elements of the first type of facial feature data to obtain flattened data; and concatenating the flattened data to generate the first feature vector.
HE discloses of flattening all elements of the first type of facial feature data to obtain flattened data(fig. 2, step S203; page 7, … the flattening layer pulls the human face feature drawing output by the main module as the one-dimensional convolution feature with length k);
and concatenating the flattened data to generate the first feature vector (fig. 2, S203; page 7, …the neck module comprises a flattening layer and a full connecting layer, wherein the flattening layer pulls the human face feature drawing output by the main module as the one-dimensional convolution feature with length k, For example, the three-dimensional or four-dimensional human face feature matrix is pulled into a one-dimensional feature vector (i.e., a convolution feature vector); then, using the full connection layer to transform the convolution characteristic into one-dimensional representation characteristic, namely using the full connection layer to transform the one-dimensional characteristic vector after leveling; in which, the full connecting layer concatenating the flattened data to generate the first feature vector). It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to incorporate the technology that flattening all elements of the first type of facial feature data to obtain flattened data; and concatenating the flattened data to generate the first feature vector as a modification to the method for the benefit of that for getting some useful characteristics (page 7).
Regarding claim 23, the combination of YOO and DU teaches the limitations recited in claim 21 as discussed above.
It is noticed that YOO does not disclose explicitly of the bitstream comprises entropy coded data of flattened data that is generated by flattening all elements of the first type of facial feature data.
HE discloses of the entropy coded data of flattened data that is generated by flattening all elements of the first type of facial feature data (fig. 2, step S203; page 7, … the flattening layer pulls the human face feature drawing output by the main module as the one-dimensional convolution feature with length k).
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to incorporate the technology that entropy coded data of flattened data that is generated by flattening all elements of the first type of facial feature data as a modification to the method for the benefit of that for getting some useful characteristics (page 7).
Regarding claim 24, the combination of YOO and DU teaches the limitations recited in claim 21 as discussed above.
It is noticed that YOO does not disclose explicitly of flattening all elements of the first type of facial feature data to obtain flattened data; and concatenating the flattened data to generate the first feature vector.
HE discloses of flattening all elements of the first type of facial feature data to obtain flattened data(fig. 2, step S203; page 7, … the flattening layer pulls the human face feature drawing output by the main module as the one-dimensional convolution feature with length k);
and concatenating the flattened data to generate the first feature vector (fig. 2, S203; page 7, …the neck module comprises a flattening layer and a full connecting layer, wherein the flattening layer pulls the human face feature drawing output by the main module as the one-dimensional convolution feature with length k, For example, the three-dimensional or four-dimensional human face feature matrix is pulled into a one-dimensional feature vector (i.e., a convolution feature vector); then, using the full connection layer to transform the convolution characteristic into one-dimensional representation characteristic, namely using the full connection layer to transform the one-dimensional characteristic vector after leveling; in which, the full connecting layer concatenating the flattened data to generate the first feature vector). It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to incorporate the technology that flattening all elements of the first type of facial feature data to obtain flattened data; and concatenating the flattened data to generate the first feature vector as a modification to the method for the benefit of that for getting some useful characteristics (page 7).
Allowable Subject Matter
11. Claim 4 would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claim.
Claim 5 and its depend claim 6 would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claim.
Claim 9 and its depend claim 10 would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claim.
Claim 11 and its depend claim 12 would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claim.
The following is a statement of reasons for the indication of allowable subject matters:
For claim 4 ,the prior art does not disclose or suggest the limitations of “wherein the first type of facial feature data comprises a plurality of arrays representing the facial picture, and the flattened data comprise a plurality of vectors respectively corresponding to the arrays.”
For claim 5 and its depend claim 6, the prior art does not disclose or suggest the limitations of “wherein the post-processing comprises: splitting the second feature vector to generate split data; and reshaping the split data to generate the second type of facial feature data.”
For claim 9 and its depend claim 10 ,the prior art does not disclose or suggest the limitations of “wherein transforming the first type of facial feature data into the second type of facial feature data comprises: concatenating flattened data to generate a first feature vector that applies to a translator, the flattened data being generated by flattening all elements of the first type of facial feature data; translating, by the translator, the first feature vector into a second feature vector being associated with the second type of facial feature data; splitting the second feature vector to generate split data; and reshaping the split data to generate the second type of facial feature data.”
For claim 11 and its depend claim 12 ,the prior art does not disclose or suggest the limitations of “wherein transforming the first type of facial feature data into the second type of facial feature data comprises: translating, by a translator, a first feature vector being associated with the first type of facial feature data into a second feature vector being associated with the second type of facial feature data, wherein the first feature vector is concatenated from flattened data that is generated by flattening all elements of the first type of facial feature data; splitting the second feature vector to generate split data; and reshaping the split data to generate the second type of facial feature data.”
12. Conclusion
. Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, 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 extension fee pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the date of this final action.
Contact Information
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ZAIHAN JIANG whose telephone number is (571)272-1399. The examiner can normally be reached on flexible.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Sath Perungavoor can be reached on (571)272-7455. The fax phone number for the organization where this application or proceeding is assigned is 571-270-0655.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/ZAIHAN JIANG/Primary Examiner, Art Unit 2488