IP Library › Granted Patent US 12,288,277
Granted Patent B2
US 12,288,277 · App. 17/827,394 · Granted Apr 29, 2025

High-precision semantic image editing using neural networks for synthetic data generation systems and applications

Inventors: Huan Ling (Toronto, CA); Karsten Kreis (Vancouver, CA); Daiqing Li (Oakville, CA); Seung Wook Kim (Toronto, CA); Antonio Torralba Barriuso (Somerville, MA); Sanja Fidler (Toronto, CA)
Assignee: NVIDIA Corporation
G06T11/60G06T7/10G06V10/774G06V10/776G06T2200/24G06T2207/20021G06T2207/20081G06T2207/20084
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,288,277
App. No.
17/827,394
Granted
Apr 29, 2025
Kind
B2
Abstract

In various examples, high-precision semantic image editing for machine learning systems and applications are described. For example, a generative adversarial network (GAN) may be used to jointly model images and their semantic segmentations based on a same underlying latent code. Image editing may be achieved by using segmentation mask modifications (e.g., provided by a user, or otherwise) to optimize the latent code to be consistent with the updated segmentation, thus effectively changing the original, e.g., RGB image. To improve efficiency of the system, and to not require optimizations for each edit on each image, editing vectors may be learned in latent space that realize the edits, and that can be directly applied on other images with or without additional optimizations. As a result, a GAN in combination with the optimization approaches described herein may simultaneously allow for high precision editing in real-time with straightforward compositionality of multiple edits.

Claims (81)

1. A processor comprising:

one or more circuits to:

generate, using a generative adversarial network (GAN) and based at least on a first point in a latent space of the GAN, a first image representing one or more objects and a segmentation mask associated with the one or more objects represented by the first image;

generate, based at least on input data representing one or more modifications to one or more first features of the one or more objects, an updated segmentation mask by at least modifying at least portions of the segmentation mask representing the one or more first features to include one or more second features;

determine a vector based at least on the updated segmentation mask and data that associates the vector with the one or more modifications to the one or more first features;

determine, based at least on the first point and the vector, a second point in the latent space different from the first point, the second point corresponding to the one or more second features; and

generate, using the GAN and based at least on the second point, a second image of the one or more objects that includes the one or more second features.

2. The processor of claim 1 , wherein the second point in the latent space is further determined using one or more latent code optimization iterations.

3. The processor of claim 2 , wherein at least one latent code optimization iteration of the one or more latent code optimization iterations includes backpropagating one or more gradients through the GAN.

4. The processor of claim 1 , wherein the second point in the latent space is determined using one or more loss functions, the one or more loss functions including at least one of:

a first loss function that penalizes one or more differences between the first image and the second image outside of an editing region corresponding to the one or more modifications;

a second loss function that rewards at least one difference between the first image and the second image inside of the editing region corresponding to the one or more modifications; or

a third loss function that penalizes one or more changes to an identity of an object between the first image and the second image.

5. The processor of claim 1 , wherein the one or more circuits are further to embed the first image into the latent space of the GAN to determine the first point.

6. The processor of claim 1 , wherein the one or more circuits are further to generate, using the GAN and based at least on the second point in the latent space of the GAN, another segmentation mask corresponding to the second image.

7. The processor of claim 1 , wherein the vector is further associated with an edit type associated with the one or more modifications to the one or more first features.

8. The processor of claim 7 , wherein the one or more circuits are further to:

determine, using the vector and based at least on a third point in the latent space of the GAN corresponding to an embedded image, a fourth point in the latent space of the GAN; and

generate, using the GAN and based at least on the fourth point in the latent space of the GAN, a third image corresponding to the embedded image with at least one edit corresponding to the edit type.

9. The processor of claim 1 , wherein the processor is comprised in at least one of:

a control system for an autonomous or semi-autonomous machine;

a perception system for an autonomous or semi-autonomous machine;

a system for performing simulation operations;

a system for performing digital twin operations;

a system for performing light transport simulation;

a system for performing collaborative content creation for 3D assets;

a system for performing deep learning operations;

a system implemented using an edge device;

a system implemented using a robot;

a system for performing conversational AI operations;

a system for generating synthetic data;

a system incorporating one or more virtual machines (VMs);

a system implemented at least partially in a data center; or

a system implemented at least partially using cloud computing resources.

10. The processor of claim 1 , wherein the modifying the one or more first features to generate the one or more second features comprises at least one of:

modifying one or more sizes of the one or more first features;

modifying one or more colors of the one or more first features;

modifying one or more shapes of the one or more first features; or

modifying one or more orientations of the one or more first features.

11. The processor of claim 1 , wherein the vector is learned using a second segmentation mask associated with one or more second objects that includes the one or more features and a second updated segmentation mask that includes the one or more modifications to the one or more features of the one or more second objects.

12. A system comprising:

one or more processing units comprising processing circuitry to:

learn an editing vector between a first point and a second point in a latent space of a generative adversarial network (GAN), the first point corresponding to a first segmentation mask representing one or more features associated with one or more first objects and the second point corresponding to a first updated segmentation mask representing the one or more features as updated using one or more modifications;

determine a third point in the latent space corresponding to an image, the images depicting one or more second objects that include the one or more features;

determine the editing vector based at least on a second segmentation mask representing the one or more features of the one or more second objects and a second updated segmentation mask representing the one or more features of the one or more second objects as updated using the one or more modifications;

determine, based at least on the editing vector and the third point in the latent space, a fourth point in the latent space that is associated with the one or more features as updated; and

generate an output image based at least on the GAN processing data corresponding to the fourth point, the output image representing the one or more second objects including the one or more features as updated.

13. The system of claim 12 , wherein the first segmentation mask is generated based at least on the GAN processing data corresponding to the first point.

14. The system of claim 12 , wherein the third point in the latent space is determined based at least on embedding the image into the latent space of the GAN.

15. The system of claim 12 , wherein the editing vector is learned by, at least in part, using one or more loss functions through one or more latent code optimization iterations.

16. The system of claim 12 , wherein the editing vector is learned by, at least in part, using one or more loss functions, the one or more loss functions including at least one of:

a first loss function that penalizes differences between a first image corresponding to the first point and a second image corresponding to the second point outside of an editing region corresponding to the one or more modifications; or

a second loss function that penalizes a lack of a difference between the first image and the second image inside of the editing region corresponding to the one or more modifications; or

a third loss function that penalizes one or more changes to an identity of the one or more first objects between the first image and the second image.

17. The system of claim 12 , wherein the system is comprised in at least one of:

a control system for an autonomous or semi-autonomous machine;

a perception system for an autonomous or semi-autonomous machine;

a system for performing simulation operations;

a system for performing digital twin operations;

a system for performing light transport simulation;

a system for performing collaborative content creation for 3D assets;

a system for performing deep learning operations;

a system implemented using an edge device;

a system implemented using a robot;

a system for performing conversational AI operations;

a system for generating synthetic data;

a system incorporating one or more virtual machines (VMs);

a system implemented at least partially in a data center; or

a system implemented at least partially using cloud computing resources.

18. The system of claim 12 , wherein:

the first point is associated with the first segmentation mask representing the one or more features associated with the one or more first objects; and

the second point is associated with the first updated segmentation mask representing the one or more features as updated.

19. The system of claim 12 , wherein the one or more processing units further comprise processing circuitry go:

store data representing the editing vector in association with the one or more modifications to the one or more features; and

wherein the determination of the editing vector is further based at least on the editing vector being associated with the one or more modifications.

20. A method comprising:

determining a first point in a latent space that is associated with a segmentation mask representing one or more objects;

generating an updated segmentation mask based at least on updating one or more first features associated with one or more objects as represented by the segmentation mask to include one or more second features;

determining, based at least on the updated segmentation mask, a vector that is associated with the updating the one or more first features to include the one or more second features;

determining, based at least on the first point and the vector, a second point in the latent space different from the first point, the second point corresponding to the one or more second features; and

generating an image using a generative adversarial network (GAN) and based at least on the second point in the latent space.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jun 6, 2022
From: LING, HUAN; KREIS, KARSTEN; LI, DAIQING; KIM, SEUNG WOOK; BARRIUSO, ANTONIO TORRALBA; FIDLER, SANJA
To: NVIDIA CORPORATION
Reel/Frame 060111/0602 →
Continuity (2)
Provisional Application 63194737 · May 28, 2021
Related Publication 20220383570A1 · Dec 1, 2022
References Cited (90)
US 20220108417A1 · Liu · 2022 [cited by examiner]
US 20220383570A1 · Ling et al. · 2022 [cited by applicant]
US 20230342986A1 · Pinto · 2023 [cited by examiner]
WO 2022251693A1 · 2022 [cited by applicant]
Lee et al., MaskGAN: Towards Diverse and Interactive Facial Image Manipulation, Jul. 27, 2019, Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR); Machine Learning (cs.LG), pp. 1-20 (Year: 2019). [cited by examiner]
Wang, et al.; “High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs”; https://arxiv.org/abs/1711.11585; Aug. 20, 2018, 14 pgs. [cited by applicant]
Luan, et al.; “Deep Photo Style Transfer”; https://arxiv.org/abs/1703.07511; Apr. 11, 2017, 9 pgs. [cited by applicant]
Liu, et al.; “Unsupervised Image-to-Image Translation Networks”; https://arxiv.org/abs/1703.00848; Jul. 23, 2018, 11 pgs. [cited by applicant]
Li, et al.; “A Closed-Form Solution to Photorealistic Image Stylization”; https://arxiv.org/abs/1802.06474; Jul. 27, 2018, 23 pgs. [cited by applicant]
Kazemi, et al.; “Style and Content Disentanglement in Generative Adversarial Networks”; https://arxiv.org/abs/1811.05621; Nov. 14, 2018, 9 pgs. [cited by applicant]
Yoo, et al.; “Photorealistic Style Transfer via Wavelet Transforms”; https://arxiv.org/abs/1903.09760; Sep. 29, 2019, 17 pgs. [cited by applicant]
Perarnau, et al.; “Invertible Conditional GANs for Image Editing”; https://arxiv.org/abs/1611.06355; Nov. 19, 2016, 9 pgs. [cited by applicant]
Donahue, et al.; “Adversarial Feature Learning”; https://arxiv.org/abs/1605.09782; Apr. 3, 2017, 18 pgs. [cited by applicant]
Brock, et al.; “Neural Photo Editing with Introspective Adversarial Networks”; https://arxiv.org/abs/1609.07093; Feb. 6, 2017, 15 pgs. [cited by applicant]
Dumoulin, et al.; “Adversarially Learned Inference”; https://arxiv.org/abs/1606.00704; Feb. 21, 2017, 18 pgs. [cited by applicant]
Richardson, et al.; “Encoding in Style: A StyleGAN Encoder for Image-To-Image Translation”; https://arxiv.org/abs/2008.00951; Apr. 21, 2021, 21 pgs. [cited by applicant]
Zhu, et al.; “Generative Visual Manipulation on the Natural Image Manifold”; https://arxiv.org/abs/1609.03552; Dec. 16, 2018, 16 pgs. [cited by applicant]
Yeh, et al.; “Semantic Image Inpainting with Deep Generative Models”; https://arxiv.org/abs/1607.07539; Jul. 13, 2017, 19 pgs. [cited by applicant]
Lipton, et al.; “Precise Recovery of Latent Vectors from Generative Adversarial Networks”; https://arxiv.org/abs/1702.04782; Feb. 17, 2017, 4 pgs. [cited by applicant]
Abdal, et al.; “Images2StyleGAN: How to Embed Images into the StyleGAN Latent Space?”; https://arxiv.org/abs/1904.03189; Sep. 2, 2019, 23 pgs. [cited by applicant]
Huh, et al.; “Transforming and Projecting Images into Class-conditional Generative Networks”; https://arxiv.org/abs/2005.01703; Aug. 27, 2020, 27 pgs. [cited by applicant]
Creswell, et al.; “Inverting the Generator of A Generative Adversarial Network”; https://arxiv.org/abs/1611.05644; Nov. 17, 2016, 9 pgs. [cited by applicant]
Raj; et al.; “GAN-Based Projector for Faster Recovery with Convergence Guarantees in Linear Inverse Problems”; In 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pp. 5601-5610, 2019; 10 pgs. [cited by applicant]
Bau, et al.; “Seeing What A Gan Cannot Generate”; In 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pp. 4501-4510, 2019, 10 pgs. [cited by applicant]
Zhu, et al.; “In-Domain GAN Inversion for Real Image Editing”; https://arxiv.org/abs/2004.00049; Jul. 16, 2020, 31 pgs. [cited by applicant]
Deng, et al.; “ArcFace: Additive Angular Margin Loss for Deep Face Recognition”; https://arxiv.org/abs/1801.07698; Sep. 4, 2022; 17 pgs. [cited by applicant]
Kingma, et al.; “Adam: A Method for Stochastic Optimization”; https://arxiv.org/abs/1412.6980; Jan. 30, 2017, 15 pgs. [cited by applicant]
Seitzer; “Pytorch-FID: FID Score for PyTorch”; https://github.com/mseitzer/pytorch-fid; Aug. 2020; Version 0.1.1.; 4 pgs. [cited by applicant]
Heusel, et al.; “GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium”; https://arxiv.org/abs/1706.08500; Jan. 12, 2018, 38 pgs. [cited by applicant]
Vaccari, et al.; “Deepfakes and Disinformation: Exploring the Impact of Synthetic Political Video on Deception, Uncertainty, and Trust in News”; Social Media + Society, Jan.-Mar. 2020; 13 pgs. [cited by applicant]
Nguyen, et al.; “Deep Learning for Deepfakes Creation and Detection: A Survey”; https://arxiv.org/abs/1909.11573; Aug. 11, 2022, 19 pgs. [cited by applicant]
Mirsky, et al.; “The Creation and Detection of Deepfakes: A Survey”; https://arxiv.org/abs/2004.11138; Sep. 13, 2020; 38 pgs. [cited by applicant]
Grover, et al.; “Bias Correction of Learned Generative Models using Likelihood-Free Importance Weighting”; https://arxiv.org/abs/1906.09531; Nov. 3, 2019, 18 pgs. [cited by applicant]
Choi, et al.; “Fair Generative Modeling via Weak Supervision”; https://arxiv.org/abs/1910.12008; Jun. 30, 2020; 22 pgs. [cited by applicant]
Yu, et al.; “Inclusive GAN: Improving Data and Minority Coverage in Generative Models”; https://arxiv.org/abs/2004.03355; Aug. 23, 2020; 22 pgs. [cited by applicant]
Lee, et al.; “Self-Diagnosing GAN: Diagnosing Underrepresented Samples in Generative Adversarial Networks”; https://arxiv.org/abs/2102.12033; Oct. 26, 2021; 34 pgs. [cited by applicant]
Ling, Huan; International Search Report and Written Opinion for PCT Patent Application No. PCT/US2022/031427, filed May 27, 2022, mailed Sep. 15, 2022, 11 pgs. [cited by applicant]
Rameen Abdal et al.; “StyleFlow: Attribute-conditioned Exploration of StyleGAN-Generated Images using Conditional Continuous Normalizing Flows”, ARXIV.org, Cornell University Library; Sep. 20, 2020, 22 pgs. [cited by applicant]
David Bau et al.: “Paint by Word”, ARXIV.org, Mar. 24, 2021, 10 pgs. [cited by applicant]
Daiqing Li et al: “Semantic Segmentation with Generative Models: Semi-Supervised Learning and Strong Out-of-Domain Generalization” ARXIV.org; Cornell University Library, Apr. 12, 2021, 12 pgs. [cited by applicant]
Jianjin Xu et al: “Linear Semantics in Generative Adversarial Networks”, ARXIV.org, Cornell University Library, Apr. 1, 2021, 23 pgs. [cited by applicant]
Karras, et al., “A Style-Based Generator Architecture for Generative Adversarial Networks”, https://arxiv.org/abs/1812.04948, Mar. 29, 2019, 12 pgs. [cited by applicant]
Zhang, et al., “The Unreasonable Effectiveness of Deep Features as a Perceptual Metric”, https://arxiv.org/abs/1801.03924, Apr. 10, 2018, 14 pgs. [cited by applicant]
Zhang, et al., “DatasetGAN: Efficient Labeled Data Factory with Minimal Human Effort”, https://arxiv.org/abs/2104.06490, Apr. 20, 2021, 11 pgs. [cited by applicant]
Bailey; “The Tools of Generative Art, from Flash to Neural Networks”, Art in America, https://www.artnews.com/art-in-america/features/generative-art-tools-flash-processing-neural-networks-1202674657/; Jan. 8, 2020, 7 pg… [cited by applicant]
Goodfellow, et al., “Generative Adversarial Nets”, In Advances in neural information processing systems, 2014, 9 pgs. [cited by applicant]
Radford, et al., “Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks”, https://arxiv.org/abs/1511.06434; Jan. 7, 2016, 16 pgs. [cited by applicant]
Kerras, et al., “Growing of Gans for Improved Quality, Stability, and Variation”, https://arxiv.org/abs/1710.10196, Feb. 26, 2018, 26 pgs. [cited by applicant]
Karras, et al., “Analyzing and Improving the Image Quality of StyleGAN”, https://arxiv.org/abs/1912.04958, Mar. 23, 2020, 21 pgs. [cited by applicant]
Choi, et al., “StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation”, https://arxiv.org/abs/1711.09020, Sep. 21, 2018, 15 pgs. [cited by applicant]
Lee, et al., “MaskGAN: Towards Diverse and Interactive Facial Image Manipulation”, https://arxiv.org/abs/1907.11922, Apr. 1, 2020, 20 pgs. [cited by applicant]
Wu, et al., “Cascade EF-GAN: Progressive Facial Expression Editing with Local Focuses”, https://arxiv.org/abs/2003.05905, Mar. 25, 2020, 18 pgs. [cited by applicant]
Shen, et al., “Interpreting the Latent Space of GANs for Semantic Face Editing”, https://arxiv.org/abs/1907.10786, Mar. 31, 2020, 12 pgs. [cited by applicant]
Shen, et al., “InterFaceGAN: Interpreting the Disentangled Face Representation Learned by GANs”, https://arxiv.org/abs/2005.09635, Oct. 29, 2020, 16 pgs. [cited by applicant]
Alharbi, et al., “Disentangled Image Generation Through Structured Noise Injection”, In Porceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun. 2020, 9 pgs. [cited by applicant]
Hou, et al., “GuidedStyle: Attribute Knowledge Guided Style Manipulation for Semantic Face Editing”, https://arxiv.org/abs/2012.11856, Dec. 22, 2020, 10 pgs. [cited by applicant]
Cherepkov, et al., “Navigating the GAN Parameter Space for Semantic Image Editing”, In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021, 10 pgs. [cited by applicant]
Collins, et al., “Editing in Style: Uncovering the Local Semantics of GANs”, https://arxiv.org/abs/2004.14367, May 21, 2020, 23 pgs. [cited by applicant]
Zhu, et al., “SEAN: Image Synthesis with Semantic Region-Adaptive Normalization”, https://arxiv.org/abs/1911.12861, May 24, 2020, 19 pgs. [cited by applicant]
Lewis, et al., “Vogue: Try-on by StyleGAN Interpolation Optimization”, arXiv preprint arXiv:2101.02285, 2021, 15 pgs. [cited by applicant]
Kim, et al., “Exploiting Spatial Dimensions of Latent in GAN for Real-time Image Editing”, https://arxiv.org/abs/2104.14754, Jun. 23, 2021, 25 pgs. [cited by applicant]
Chen, et al., “DeepFaceDrawing: Deep Generation of Face Images from Sketches”, ACM Trans. Graph., Art. 39, Jun. 5, 2020, 16 pgs. [cited by applicant]
He, et al., “AttGAN: Facial Attribute Editing by Only Changing What You Want”, https://arxiv.org/abs/1711.10678, Jul. 25, 2018, 16 pgs. [cited by applicant]
Bau, et al., “GAN Dissection: Visualizing and Understanding Generative Adversarial Networks”, https://arxiv.org/abs/1811.10597, Dec. 8, 2018, 18 pgs. [cited by applicant]
Bau, et al., “Semantic Photo Manipulation with a Generative Image Prior”, https://arxiv.org/abs/2005.07727, Sep. 12, 2020, 11 pgs. [cited by applicant]
Plumerault, et al., “Controlling Generative Models with Continuous Factors of Variations”, https://arxiv.org/abs/2001.10238, Jan. 28, 2020, 17 pgs. [cited by applicant]
Harkonen, et al., “GANSpace: Discovering Interpretable GAN Controls”, https://arxiv.org/abs/2004.02546, Dec. 14, 2020, 29 pgs. [cited by applicant]
Bau, et al., “Rewriting A Deep Generative Model”, https://arxiv.org/abs/2007.15646, Jul. 30, 2020, 31 pgs. [cited by applicant]
Efros, et al., “Image Quilting for Texture Synthesis and Transfer”, SIGGRAPH '01, p. 341-346, New York, NY USA, 2001. Association for Computing Machinery, 6 pgs. [cited by applicant]
Hertzmann, et al., “Image Analogies”, https://www.researchgate.net/publication/2406594_Image_Analogies, 2001, 14 pgs. [cited by applicant]
Reinhard, et al., “Color Transfer Between Images”, IEEE Computer Graphics and Applications, vol. 21, 2001, 9 pgs. [cited by applicant]
Perez, et al., “Poisson Image Editing”, SIGGRAPH '03, p. 313-318, New York, NY, USA, 2003. Association for Computing Machinery. 6 pgs. [cited by applicant]
Schaefer, et al., “Image Deformation Using Moving Least Squares”, ACM Trans. Graph., Art. 25, 2006, 8 pgs. [cited by applicant]
Barnes, et al., “Patchmatch: A Randomized Correspondence Algorithm for Structural Image Editing”, ACM Trans. Graph., Art. 28, 2009, 10 pgs. [cited by applicant]
Tao, et al., “Error-tolerant Image Compositing”, in ECCV, 2010, 14 pgs. [cited by applicant]
Gatys, et al., “Image Style Transfer Using Convolutional Neural Networks”, in 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, 10 pgs. [cited by applicant]
Zhu, et al., “Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks”, https://arxiv.org/abs/1703.10593, Aug. 27, 2020, 18 pgs. [cited by applicant]
Portenier, et al., “FaceShop:Deep Sketch-based Face Image Editing”, https://arxiv.org/abs/1804.08972, Jun. 7, 2018, 13 pgs. [cited by applicant]
Ling, et al., “Variational Amodal Object Completion”, Advances in Neural Information Processing Systems, 2020, 12 pgs. [cited by applicant]
Park, et al., “Swapping Autoencoder for Deep Image Manipulation”, https://arxiv.org/abs/2007.00653, Dec. 14, 2020, 23 pgs. [cited by applicant]
Kim, et al.; “DriveGAN: Towards a Controllable High-Quality Neural Simulation”, https://arxiv.org/abs/2104.15060; Apr. 30, 2021, 16 pgs. [cited by applicant]
Kingma, et al.; “Auto-Encoding Variational Bayes”, https://arxiv.org/abs/1312.6114; May 1, 2014, 14 pgs. [cited by applicant]
Rezende, et al.; “Stochastic Backpropogation and Approximate Inference in Deep Generative Models”, https://arxiv.org/abs/1401.4082; May 30, 2014, 14 pgs. [cited by applicant]
Brock, et al.; “Large Scale GAN Training for High Fidelity Natural Image Synthesis”, https://arxiv.org/abs/1809.11096; Feb. 25, 2019, 35 pgs. [cited by applicant]
Park, et al.: “Semantic Image Synthesis with Spatially-Adaptive Normalization”, https://arxiv.org/abs/1903.07291; Nov. 5, 2019, 19 pgs. [cited by applicant]
Goetschalckx, et al.; “GANalyze: Toward Visual Definitions of Cognitive Image Properties”, https://arxiv.org/abs/1906.10112; Jun. 24, 2019, 17 pgs. [cited by applicant]
Jahanian, et al.; “On the ‘Steerability’ of Generative Adversarial Networks”; https://arxiv.org/abs/1907.07171, Feb. 17, 2020, 31 pgs. [cited by applicant]
Voynov, et al.; “Unsupervised Discovery of Interpretable Directions in the GAN Latent Space”; https://arxiv.org/abs/2002.03754; Jun. 24, 2020, 15 pgs. [cited by applicant]
Shen, et al.; “Closed-Form Factorization of Latent Semantics in GANs”; https://arxiv.org/abs/2007.06600; Apr. 3, 2021, 9 pgs. [cited by applicant]
Ling, Huan; International Preliminary Report on Patentability for PCT Application No. PCT/US2022/031427, filed May 27, 2022, mailed Dec. 7, 2023, 8 pgs. [cited by applicant]
Cited By (1)
US 12,541,818