Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
Applicant’s arguments filed 02/17/2026 on pages 9-11 of Remarks regarding the rejection under 35 U.S.C. 103 with respect to claims 1-18 and 25-30 have been fully considered but are moot in view of the new combination of references.
Claim Objections
Claim 25 is objected to because of the following informalities:
In claim 25, line 1, “A image generation system” should read, “An image generation system”.
In claim 25, line 3, “using a attribute classifier” should read, “using an attribute classifier”.
Appropriate correction is required.
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, 2, 7, 8, 13, 14, 25 and 26 are rejected under 35 U.S.C. 103 as being unpatentable over Ju et al. (Adaptive semantic attribute decoupling for precise face image editing); hereinafter Ju in view of Pang et al. (Learning Latent Space Energy-Based Prior Model); hereinafter Pang
Claim 1 is rejected over Ju and Pang.
Regarding claim 1, Ju teaches a processor, comprising:
one or more circuits to:
classify, using an attribute classifier based on [one or more energy-based models (EBMs)] formulated in a latent space of] one or more generative neural networks, (Ju [page 2910, 3.1 Attribute control vector adaptive to individual face]: “in combination of a decoder, a pre-trained generator, and a attribute classifier.”; and [page 11, 3.1.2]: “the pre-trained generator of StyleGAN [22] model is employed, which is able to output 1024 × 1024 face images. The input of this generator can be 512-dimensional noise vector, or the vector after decoupling the noise. It is found through experiment that the decoupled input vector could more purify control the generating features”)
one or more specified attributes of one or more dynamically-configurable attributes of one or more objects to a distribution conditioned on the one or more specified attributes, (Ju [page 2910]: “For any user specified attribute, control vector optimal to every face image is extracted from the latent code through a proposed decoder”; [page 2911, Eq (4)]: “C(n) is used to train control vector the ability of changing attributes”)
sample, [based on the EBM formulation,] the classified one or more specified attributes according to the distribution from the attribute classifier in the latent space of the one or more generative neural networks to provide one or more latent space features for the one or more specified attributes to the one or more generative neural networks; and (Ju [page 2910]: “1. Random select a latent code z which correspondence to a face image; and [page 2913]: “100 latent codes are randomly selected, and Eq. 1 is used to modify these latent codes to create new faces given in Fig. 8. As can be inferred from the results, the trained control vector could add attribute to the faces and deepen it smoothly conforms to human perception.”)
generate, using the one or more generative neural networks, the one or more images depicting the one or more objects based, at least in part, on the latent space features for the one or more specified attributes. (Ju [page 11, 3.1.2]: “the pre-trained generator of StyleGAN
[22] model is employed, which is able to output 1024 × 1024 face images. The input of this generator can be 512-dimensional noise vector, or the vector after decoupling the noise. It is found through experiment that the decoupled input vector could more purify control the generating features”; and [page 2910]: “By changing the original latent code along the direction of control vector with adjustable amount and regenerate the image with the modified latent code, precise face attribute editing is realized.”)
Ju does not appear to explicitly teach based on or more energy-based models (EBMs) formulated in a latent space of
based on the EBM formulation,
However, Pang teaches based on or more energy-based models (EBMs) formulated in a latent space of
based on the EBM formulation, (“We propose to learn energy-based model (EBM) in the latent space of a generator model, so that the EBM serves as a prior model that stands on the top-down network of the generator model.”; page 1, Abstract)
It would have been obvious before the effective filing date to combine the attribute classifier of Ju with the EBM in latent space of Pang to effectively capture regularities (Pang, page 1, I Introduction). Ju and Pang are analogous art because they both concern image generation from latent space.
Claim 2 is rejected over Ju and Pang with the incorporation of claim 1.
Regarding claim 2, Ju teaches wherein the one or more images are generated using one or more unconditional generative neural networks that are not trained on the dynamically-configurable attributes. (Ju [3.1.2. Selection of pre-trained generator]: “In our framework, the pre-trained generator of StyleGAN [22] model is employed, which is able to output 1024 × 1024 face images. The input of this generator can be 512-dimensional noise vector, or the vector after decoupling the noise. It is found through experiment that the decoupled input vector could more purify control the generating features, it is used in this work.”; Note: The pre-trained generator is not trained on attributes as there is a separate attribute classifier mentioned in 3.1.3 meant to provide guidance to the decoder).
Claim 7 is claim 1 in the form of a system and is rejected for the same reasons as claim 1 stated above.
Dependent claim 8 is claim 2 in the form of a system and is rejected for the same reasons as claim 2 stated above. For the rejection of the limitations specifically pertaining to the system of claim 7, see the rejection of claim 7 above.
Claim 13 is claim 1 in the form of a method and is rejected for the same reasons as claim 1 stated above.
Dependent claim 14 is claim 2 in the form of a method and is rejected for the same reasons as claim 2 stated above. For the rejection of the limitations specifically pertaining to the method of claim 13, see the rejection of claim 13 above.
Claim 25 is rejected over Ju and Pang.
Regarding claim 25, Ju teaches a image generation system, comprising:
one or more processors to generate one or more images comprising one or more objects based, at least in part, on one or more dynamically configurable attributes of the one or objects; and
memory for storing network parameters for the one or more neural networks. (See Abstract to see that Ju recites a computer based system for facial attribute modification on an image.)
The remainder of claim 25 is claim 1 in the form of a system and is rejected for the same reasons as claim 1 stated above.
Dependent claim 26 is claim 2 in the form of a system and is rejected for the same reasons as claim 2 stated above. For the rejection of the limitations specifically pertaining to the system claim of 25, see the rejection of claim 25 above.
Claims 3, 4, 5, 6, 9, 10, 11, 12 ,15, 16, 17, 18, 27, 28, 29 and 30 are rejected under 35 U.S.C. 103 as being unpatentable over Ju and Pang in further view of Du et al. (Compositional Visual Generation with Energy Based Models); hereinafter Du
Claim 3 is rejected over Ju, Pang and Du with the incorporation of claim 1.
Regarding claim 3, Ju does not appear to explicitly teach train one or more energy-based models (EBMs) on the one or more dynamically-configurable attributes; and
generate, using the one or more energy-based models, a distribution over feature data and the one or more dynamically-configurable attributes in an image space.
However, Du teaches train one or more energy-based models (EBMs) on the one or more dynamically-configurable attributes; and (Du [page 3, 3.2 Composition of Energy-Based Models]: “We next present different ways that EBMs can compose. We consider a set of independently trained EBMs, E(x|c1), E(x|c2), . . . , E(x|cn), which are learned conditional distributions on underlying concept codes ci. Latent codes we consider include position, size, color, gender, hair style, and age, which we also refer to as concepts. Figure 2 shows three concepts and their combinations on the CelebA face dataset and attributes. In concept conjunction, given separate independent concepts (such as a particular gender, hair style, or facial expression), we wish to construct an output with the specified gender, hair style, and facial expression – the combination of each concept.”; Note: The latent codes from the distribution are the features)
generate, using the one or more energy-based models, a distribution over feature data and the one or more dynamically-configurable attributes in an image space. (Du [page 3, 3.2 Composition of Energy-Based Models]: “Latent codes we consider include position, size, color, gender, hair style, and age, which we also refer to as concepts. Figure 2 shows three concepts and their combinations on the CelebA face dataset and attributes. In concept conjunction, given separate independent concepts (such as a particular gender, hair style, or facial expression), we wish to construct an output with the specified gender, hair style, and facial expression – the combination of each concept.”; Note: The latent codes from the distribution are the encoded features)
It would have been obvious before the effective filing date to combine the attribute classifier of Ju with the energy-based model of Du to effectively configure attributes of an image (Du, page 3 paragraphs 1-2). Ju and Du are analogous art because they both concern attribute configuration in image generation.
Claim 4 is rejected over Ju, Pang and Du with the incorporation of claim 1.
Regarding claim 4, Ju does not appear to explicitly teach wherein the one or more circuits are further to transfer the distribution from the image space to a latent space to be input to one or more generative neural networks for generating the one or more images.
However, Pang teaches wherein the one or more circuits are further to transfer the distribution from the image space to a latent space to be input to one or more generative neural networks for generating the one or more images. (Pang [page 9]: “From data space EBM to latent space EBM. EBM learned in data space such as image space [53, 47, 81, 18, 27, 55, 16] can be highly multi-modal, and MCMC sampling can be difficult. We can introduce latent variables and learn an EBM in latent space, while also learning a mapping from
the latent space to the data space.”)
It would have been obvious before the effective filing date to combine the attribute classifier of Ju with the EBM in latent space of Pang to effectively capture regularities (Pang, page 1, I Introduction). Ju and Pang are analogous art because they both concern image generation from latent space.
Claim 5 is rejected over Ju, Pang and Du with the incorporation of claim 1.
Regarding claim 5, Ju does not appear to explicitly teach wherein the one or more circuits are further to utilize an ordinary differential equation (ODE) for sampling a region of the latent space corresponding to the distribution.
However, Du teaches wherein the one or more circuits are further to utilize an ordinary differential equation (ODE) for sampling a region of the latent space corresponding to the distribution. (See page 3, 3.1 Energy Based Models of Du to see that the differential equation (3) is used to sample from distributions and generate multiple different compositions of distributions).
It would have been obvious before the effective filing date to combine the attribute classifier of Ju with the energy-based model of Du to effectively configure attributes of an image (Du, page 3 paragraphs 1-2). Ju and Du are analogous art because they both concern attribute configuration in image generation.
Claim 6 is rejected over Ju, Pang and Du with the incorporation of claim 1.
Regarding claim 6, Ju teaches wherein the one or more circuits are further to generate the one or more images based, at least in part, upon features sampled from the region of the latent space corresponding to the distribution. (Ju [page 11, 3.1.2]: “the pre-trained generator of StyleGAN [22] model is employed, which is able to output 1024 × 1024 face images. The input of this generator can be 512-dimensional noise vector, or the vector after decoupling the noise. It is found through experiment that the decoupled input vector could more purify control the generating features”; [page 2910]: “By changing the original latent code along the direction of control vector with adjustable amount and regenerate the image with the modified latent code, precise face attribute editing is realized.”; [page 2910]: “1. Random select a latent code z which correspondence to a face image; and [page 2913]: “100 latent codes are randomly selected, and Eq. 1 is used to modify these latent codes to create new faces given in Fig. 8. As can be inferred from the results, the trained control vector could add attribute to the faces and deepen it smoothly conforms to human perception.”))
Dependent claim 9 is claim 3 in the form of a system and is rejected for the same reasons as claim 3 stated above. For the rejection of the limitations specifically pertaining to the system of claim 7, see the rejection of claim 7 above.
Dependent claim 10 is claim 4 in the form of a system and is rejected for the same reasons as claim 4 stated above. For the rejection of the limitations specifically pertaining to the system of claim 7, see the rejection of claim 7 above.
Dependent claim 11 is claim 5 in the form of a system and is rejected for the same reasons as claim 5 stated above. For the rejection of the limitations specifically pertaining to the system of claim 7, see the rejection of claim 7 above.
Dependent claim 12 is claim 6 in the form of a system and is rejected for the same reasons as claim 6 stated above. For the rejection of the limitations specifically pertaining to the system of claim 7, see the rejection of claim 7 above.
Dependent claim 15 is claim 3 in the form of a method and is rejected for the same reasons as claim 3 stated above. For the rejection of the limitations specifically pertaining to the method of claim 13, see the rejection of claim 13 above.
Dependent claim 16 is claim 4 in the form of a method and is rejected for the same reasons as claim 4 stated above. For the rejection of the limitations specifically pertaining to the method of claim 13, see the rejection of claim 13 above.
Dependent claim 17 is claim 5 in the form of a method and is rejected for the same reasons as claim 5 stated above. For the rejection of the limitations specifically pertaining to the method of claim 13, see the rejection of claim 13 above.
Dependent claim 18 is claim 6 in the form of a method and is rejected for the same reasons as claim 6 stated above. For the rejection of the limitations specifically pertaining to the method of claim 13, see the rejection of claim 13 above.
Dependent claim 27 is claim 3 in the form of a system and is rejected for the same reasons as claim 3 stated above. For the rejection of the limitations specifically pertaining to the system claim of 25, see the rejection of claim 25 above.
Dependent claim 28 is claim 4 in the form of a system and is rejected for the same reasons as claim 4 stated above. For the rejection of the limitations specifically pertaining to the system claim of 25, see the rejection of claim 25 above.
Dependent claim 29 is claim 5 in the form of a system and is rejected for the same reasons as claim 5 stated above. For the rejection of the limitations specifically pertaining to the system claim of 25, see the rejection of claim 25 above.
Dependent claim 30 is claim 6 in the form of a system and is rejected for the same reasons as claim 6 stated above. For the rejection of the limitations specifically pertaining to the system claim of 25, see the rejection of claim 25 above.
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 nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DAVID H TRAN whose telephone number is (703)756-1525. The examiner can normally be reached M-F 9:30 am - 5:30 pm.
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, Viker Lamardo can be reached at (571) 270-5871. 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.
/DAVID H TRAN/Examiner, Art Unit 2147
/VIKER A LAMARDO/Supervisory Patent Examiner, Art Unit 2147
Prosecution Insights