IP Library › Granted Patent US 12,632,998
Granted Patent B2
US 12,632,998 · App. 17/741,959 · Granted May 19, 2026

Plausible dayscale timelapse generation method and computing device

Inventors: Gleb Mikhailovich Sterkin (Moscow, RU); Ivan Aleksandrovich Anokhin (Moscow, RU); Taras Andreevich Khakhulin (Moscow, RU); Aleksei Vladislavovich Kharlamov (Moscow, RU); Denis Mikhailovich Korzhenkov (Moscow, RU); Victor Sergeevich Lempitsky (Moscow, RU); Sergey Igorevich Nikolenko (Moscow, RU); Aleksei Sergeevich Silvestrov (Moscow, RU); Pavel Ilich Solovev (Moscow, RU)
Assignee: Samsung Electronics Co., Ltd.
G06T11/00G06N20/20G06T3/40G06T7/10G06T2207/10016G06T2207/20016G06T2207/20081G06T2207/20084G06T2207/20132G06T2207/20221G06T2210/22
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 12,632,998
App. No.
17/741,959
Granted
May 19, 2026
Kind
B2
Abstract

The disclosure relates to a field of plausible timelapse image(s) generation from a single image. A method of generating one or more images of a plausible dayscale timelapse sequence based on a content image using a trained generative neural network and a trained merging neural network is provided. The method includes receiving the content image and one of one or more predefined styles respectively corresponding to times of day to be applied to the content image or style images having styles to be applied to the content image, slicing the content image into n image crops, applying the trained generative neural network with each style to n image crops to obtain n image crops re-stylized according to each style, and merging the re-stylized n image crops for each style with the trained merging neural network to obtain images of a plausible dayscale timelapse sequence for the content image.

Claims (76)

1 . A method performed by a computing device of generating one or more images of a plausible dayscale timelapse sequence based on a content image using a trained generative neural network and a trained merging neural network, the method comprising:

receiving, by the computing device, the content image and one of a predefined one or more styles to be applied to the content image or one or more style images having one or more styles to be applied to the content image;

slicing, by the computing device, the content image into n image crops by slicing each image of a set of training images into n strongly overlapping image crops in a predefined manner defined by an offset direction and a stride of k pixels;

applying, by the computing device, the trained generative neural network with each of one or more styles to n image crops to obtain n image crops re-stylized according to each of one or more styles; and

merging, by the computing device, the re-stylized n image crops for each of the one or more styles with the trained merging neural network to obtain one or more images of a plausible dayscale timelapse sequence for the content image,

wherein the re-stylized n image crops are generated without directly feeding a discriminator to the generative neural network,

wherein the trained generative neural network is initially and iteratively trained in a swap mode using the following operations repeatedly performed at a number of iterations:

inputting a first image as a training example into a content encoder to obtain a content tensor being a spatially smaller representation of content in the first image;

inputting a second image as a training example into a style encoder to obtain a style vector indicative of a style in the second image;

inputting the content tensor, the style vector into a decoder to obtain, using adaptive instance normalization, a third image having content of the first image and a style of the second image, and a segmentation mask corresponding to the first image;

alternately inputting the second image and the style vector or the style vector and the third image into a conditional discriminator configured to determine when an image is plausible enough under a given style; and

alternately updating parameters of the trained generative neural network and the conditional discriminator at different iterations based on a result of a determination by the conditional discriminator to ensure training of the trained generative neural network and conditional discriminator in an adversarial manner,

wherein in updating parameters of the trained generative neural network the segmentation mask is further considered, and

wherein the first image and the second image have a first resolution.

2 . The method of claim 1 , wherein the trained generative neural network is initially and iteratively trained in one or in a combination of the following modes: a swap mode, a random mode, or an autoencoder mode.

3 . The method of claim 2 ,

wherein the trained generative neural network is initially and iteratively trained in the random mode using the following operations repeatedly performed at a number of iterations:

inputting a first image as a training example into the content encoder to obtain a content tensor being a spatially smaller representation of content in the first image;

generating a random style vector from prior distribution;

inputting the content tensor, the random style vector into the decoder to obtain, using adaptive instance normalization, a third image having a content of the first image and a style defined by the random style vector, and a segmentation mask corresponding to the first image;

inputting the random style vector and the third image into a conditional discriminator configured to determine when the third image is plausible enough under a given style; and

updating parameters of the trained generative neural network based on a result of a determination by the conditional discriminator to ensure training of the trained generative neural network and conditional discriminator in an adversarial manner, and

wherein in updating parameters of the trained generative neural network the segmentation mask is further considered.

4 . The method of claim 2 ,

wherein the trained generative neural network is initially and iteratively trained in the autoencoder mode using the following operations repeatedly performed at a number of iterations:

inputting a first image as a training example into the content encoder to obtain a content tensor being a spatially smaller representation of content in the first image;

inputting the first image as a training example into the style encoder to obtain a style vector indicative of a style in the first image;

inputting the content tensor, the style vector into the decoder to obtain, using adaptive instance normalization, a third image having the content and a style of the first image, and a segmentation mask corresponding to the first image;

comparing the first image and the third image; and

updating parameters of the trained generative neural network based on a result of the comparison, and

wherein in updating parameters of the trained generative neural network the segmentation mask is further considered.

5 . The method of claim 4 , wherein the trained generative neural network comprises one or more dense skip connections configured to convey high frequency features not conveyed by the content tensor of the first image to the decoder.

6 . The method of claim 1 ,

wherein the merging neural network is initially and iteratively trained using the following operations repeatedly performed at a number of iterations:

obtaining a set of training images each having a second resolution

downsampling each image crop of the n image crops to a first resolution;

applying the trained generative neural network in an autoencoder mode to each image crop of the n image crops to obtain n translated crops each capturing artifacts and discrepancies between a corresponding image crop of an original training image and the translated image crop produced by the trained generative neural network;

inputting the n translated crops into the merging neural network to obtain a merged image with reduced artifacts and discrepancies, comparing the merged image with a corresponding original image of the set of training images; and

updating parameters of the merging neural network based on a result of the comparison, and

wherein the merged image has the second resolution.

7 . The method of claim 1 , wherein one or more styles respectively correspond to one or more times of day to be applied to the content image.

8 . A computing device comprising:

memory storing one or more computer programs; and

one or more processors communicatively coupled to the memory,

wherein the one or more computer programs include computer-executable instructions that, when executed by the one or more processors individually or collectively, cause the computing device to:

receive a content image and one of a predefined one or more styles to be applied to the content image or one or more style images having one or more styles to be applied to the content image,

slice the content image into n image crops by slicing each image of a set of training images into n strongly overlapping image crops in a predefined manner defined by an offset direction and a stride of k pixels,

apply a trained generative neural network with each of one or more styles to n image crops to obtain n image crops re-stylized according to each of the one or more styles, and

merge the re-stylized n image crops for each of one or more styles with the trained merging neural network to obtain one or more images of a plausible dayscale timelapse sequence for the content image, and

wherein the re-stylized n image crops are generated without directly feeding a discriminator to the generative neural network,

wherein the trained generative neural network is initially and iteratively trained in a swap mode using the following operations repeatedly performed at a number of iterations:

inputting a first image as a training example into a content encoder to obtain a content tensor being a spatially smaller representation of content in the first image;

inputting a second image as a training example into a style encoder to obtain a style vector indicative of a style in the second image;

inputting the content tensor, the style vector into a decoder to obtain, using adaptive instance normalization, a third image having content of the first image and a style of the second image, and a segmentation mask corresponding to the first image;

alternately inputting the second image and the style vector or the style vector and the third image into a conditional discriminator configured to determine when an image is plausible enough under a given style; and

alternately updating parameters of the trained generative neural network and the conditional discriminator at different iterations based on a result of a determination by the conditional discriminator to ensure training of the trained generative neural network and conditional discriminator in an adversarial manner,

wherein in updating parameters of the trained generative neural network the segmentation mask is further considered, and

wherein the first image and the second image have a first resolution.

9 . A method performed by a computing device of generating one or more images of a plausible dayscale timelapse sequence based on a content image using a trained generative neural network, the method comprising:

receiving, by the computing device, the content image and at least one predefined one or more styles to be applied to the content image or one or more style images having one or more styles to be applied to the content image;

reducing, by the computing device, a resolution of the content image to a lower resolution while maintaining an aspect ratio of the content image;

applying, by the computing device, the trained generative neural network with each of one or more styles to the reduced resolution of the content image to obtain one or more reduced content images re-stylized according to each of one or more styles;

decomposing, by the computing device, each of a re-stylized content image into high-frequency components and a low-frequency component having the lower resolution while maintaining the aspect ratio;

filtering, by the computing device, the low-frequency component by considering the content of a corresponding re-stylized content image; and

generating, by the computing device, one or more images of the plausible dayscale timelapse sequence based on the filtered low-frequency component and the high-frequency components of each of corresponding re-stylized content images,

wherein the obtained one or more reduced content images re-stylized according to each of one or more styles is obtained without directly feeding a discriminator to the generative neural network, and

wherein the generated one or more images of the plausible dayscale timelapse sequence are generated by applying a differentiated guided filter to the content image and a corresponding one of the obtained one or more reduced content images re-stylized according to each of one or more styles.

10 . The method of claim 9 , wherein a Laplacian pyramid is used for the decomposition.

11 . The method of claim 9 ,

wherein a guided filter is used for the filtering, and

wherein the content of the corresponding re-stylized content image is used as a guide for the filtering.

12 . The method of claim 9 , wherein the trained generative neural network is initially and repeatedly trained in one or in a combination of the following modes: a swap mode, a random mode, or an autoencoder mode.

13 . The method of claim 9 , wherein the trained generative neural network comprises a content encoder, a style encoder, and a decoder.

14 . The method of claim 13 , wherein the content encoder comprises two downsamplings and four residual blocks after each downsampling.

15 . The method of claim 13 , wherein the style encoder contains four downsampling blocks, and a downsampled result is averaged with respect to spatial information into a three-dimensional vector.

16 . The method of claim 13 , wherein the decoder comprises five residual blocks with an adaptive instance normalizations (AdaIN) therein and two upsampling blocks.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded May 11, 2022
From: STERKIN, GLEB MIKHAILOVICH; ANOKHIN, IVAN ALEKSANDROVICH; KHAKHULIN, TARAS ANDREEVICH; KHARLAMOV, ALEKSEI VLADISLAVOVICH; KORZHENKOV, DENIS MIKHAILOVICH; LEMPITSKY, VICTOR SERGEEVICH; NIKOLENKO, SERGEY IGOREVICH; SILVESTROV, ALEKSEI SERGEEVICH; SOLOVEV, PAVEL ILICH
To: SAMSUNG ELECTRONICS CO., LTD.
Reel/Frame 059894/0156 →
Priority Claims (2)
RU RU2019136769 · Nov 15, 2019 · national
RU RU2020113037 · Apr 7, 2020 · national
Continuity (2)
Continuation PCTKR2020015687 · Nov 10, 2020
Related Publication 20220270304A1 · Aug 25, 2022
References Cited (77)
US 7365744B2 · Fang et al. · 2008 [cited by applicant]
US 8515270B1 · Posehn · 2013 [cited by applicant]
US 10535164B2 · Shlens et al. · 2020 [cited by applicant]
US 20060209090A1 · Kelly et al. · 2006 [cited by applicant]
US 20120070098A1 · Gohshi · 2012 [cited by applicant]
US 20160343402A1 · Singhal et al. · 2016 [cited by applicant]
US 20170054936A1 · Hanayama · 2017 [cited by applicant]
US 20170301095A1 · Zhang · 2017 [cited by examiner]
US 20180068463A1 · Risser · 2018 [cited by applicant]
US 20180075581A1 · Shi et al. · 2018 [cited by applicant]
US 20180322662A1 · Hellier et al. · 2018 [cited by applicant]
US 20180357800A1 · Oxholm et al. · 2018 [cited by applicant]
US 20190026870A1 · Hu et al. · 2019 [cited by applicant]
US 20190236814A1 · Shlens et al. · 2019 [cited by applicant]
US 20190244060A1 · Dundar et al. · 2019 [cited by applicant]
CN 106651766A · 2017 [cited by applicant]
CN 108470320A · 2018 [cited by applicant]
CN 109840575A · 2019 [cited by applicant]
CN 109863537A · 2019 [cited by applicant]
CN 110222722A · 2019 [cited by applicant]
WO 2017111302A1 · 2017 [cited by applicant]
Nam, S., Ma, C., Chai, M., Brendel, W., Xu, N., & Kim, S. J. (Apr. 1, 2019). End-to-end time-lapse video synthesis from a single outdoor image. arXiv.org. https://arxiv.org/abs/1904.00680 (Year: 2019). [cited by examiner]
Huang, Z., Zhang, J., & Liao, J. (Oct. 29, 2019). Style mixer: Semantic-aware multi-style transfer network. arXiv.org. https://arxiv.org/abs/1910.13093 (Year: 2019). [cited by examiner]
Dertat, A. (Oct. 8, 2017). Applied deep learning—part 3: Autoencoders. Medium. https://towardsdatascience.com/applied-deep-learning-part-3-autoencoders-1c083af4d798 (Year: 2017). [cited by examiner]
He, K., Zhang, X., Ren, S., & Sun, J. (Dec. 10, 2015). Deep residual learning for image recognition. arXiv.org. https://arxiv.org/abs/1512.03385 (Year: 2015). [cited by examiner]
Xu, X., Chen, Y.-C., & Jia, J. (Oct. 1, 2019). View independent generative Adversarial Network for . . . https://openaccess.thecvf.com/content_ICCV_2019/papers/Xu_View_Independent_Generative_Adversarial_Network_for_Nove… [cited by examiner]
Du, Z., Yang, J., Ou, C., & Zhang, T. (2019). Smallholder Crop Area Mapped with a Semantic Segmentation Deep Learning Method Remote Sensing, 11(7), 888. https://doi.org/10.3390/rs11070888 (Year: 2019). [cited by examiner]
Sheng, L., Lin, Z., Shao, J., & Wang, X. (2018). Avatar-Net: Multi-scale Zero-shot Style Transfer by Feature Decoration. ArXiv.org. https://arxiv.org/abs/1805.03857 (Year: 2018). [cited by examiner]
Park, D. Y., & Lee, K. H. (2018). Arbitrary Style Transfer with Style-Attentional Networks. ArXiv.org. https://arxiv.org/abs/1812.02342 (Year: 2018). [cited by examiner]
Shih, Y., Paris, S., Durand, F., & Freeman, W. T. (2013). Data-driven hallucination of different times of day from a single outdoor photo. ACM Transactions on Graphics, 32(6), 1-11. https://doi.org/10.1145/2508363.25084… [cited by examiner]
Xu, Z., Wilber, M., Fang, C., Hertzmann, A., & Jin, H. (2019). Learning from Multi-domain Artistic Images for Arbitrary Style Transfer. ArXiv:1805.09987 [Cs]. https://arxiv.org/abs/1805.09987 (Year: 2019). [cited by examiner]
Yoo, J., Uh, Y., Chun, S., Kang, B., & Ha, J.-W. (2019). Photorealistic Style Transfer via Wavelet Transforms. Thecvf.com, 9036-9045. https://openaccess.thecvf.com/content_ICCV_2019/html/Yoo_Photorealistic_Style_Transfe… [cited by examiner]
Huang, X., & Belongie, S. (2017). Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization. ArXiv:1703.06868 [Cs] . https://arxiv.org/abs/1703.06868 (Year: 2017). [cited by examiner]
Sveinn Palsson, Eirikur Agustsson, Radu Timofte, & Gool, L. V. (2018). Generative Adversarial Style Transfer Networks for Face Aging. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)… [cited by examiner]
Ioannou, E., & Maddock, S. (2022). Depth-aware Neural Style Transfer using Instance Normalization. White Rose Research Online (University of Leeds, the University of Sheffield, University of York). https://doi.org/10.23… [cited by examiner]
European Search Report dated Feb. 10, 2023, issued in European Patent Application No. 20886588.1. [cited by applicant]
Yichang Shih et al., Data-driven Hallucination of Different Times of Day from a Single Outdoor Photo (Nov. 1, 2013). [cited by applicant]
Zixuan Huang et al., Style Mixer: Semantic-aware Multi-Style Transfer Network (Oct. 29, 2019). [cited by applicant]
Ming-Yu Liu et al., Few-Shot Unsupervised Image-to-Image Translation (Sep. 9, 2019). [cited by applicant]
Jaejun Yoo et al., Photorealistic Style Transfer via Wavelet Transforms (Sep. 9, 2019). [cited by applicant]
Nam et al., End-to-End Time-Lapse Video Synthesis from a Single Outdoor Image (Apr. 1, 2019). [cited by applicant]
Xiong et al., Learning to Generate Time-Lapse Videos Using Multi-Stage Dynamic Generative Adversarial Networks (Sep. 22, 2019). [cited by applicant]
Pan et al., Video Generation from Single Semantic Label Map (Mar. 11, 2019). [cited by applicant]
Anoosheh et al., Night-to-Day Image Translation for Retrieval-based Localization (Sep. 26, 2018). [cited by applicant]
Choi et al., StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation (Jun. 2018). [cited by applicant]
Chu et al., CycleGAN, a Master of Steganography (Dec. 16, 2017). [cited by applicant]
Endo et al., Animating Landscape: Self-Supervised Learning of Decoupled Motion and Appearance for Single-Image Video Synthesis (Oct. 16, 2019). [cited by applicant]
He et al., Guided Image Filtering (Jun. 2013). [cited by applicant]
He et al., Progressive Color Transfer with Dense Semantic Correspondences (Dec. 12, 2018). [cited by applicant]
Huang et al., Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization (Jul. 30, 2017). [cited by applicant]
Huang et al., Multimodal Unsupervised Image-to-Image Translation (Aug. 14, 2018). [cited by applicant]
Isola et al., Image-to-Image Translation with Conditional Adversarial Networks (Nov. 21, 2016). [cited by applicant]
Kotovenko et al., Content and Style Disentanglement for Artistic Style Transfer (Oct. 2019). [cited by applicant]
Laffont et al., Transient Attributes for High-Level Understanding and Editing of Outdoor Scenes (Jul. 2014). [cited by applicant]
Lee et al., DRIT++: Diverse Image-to-Image Translation via Disentangled Representations (May 2, 2019). [cited by applicant]
Lee et al., Diverse Image-to-Image Translation via Disentangled Representations (Aug. 2, 2018). [cited by applicant]
Li et al., A Closed-form Solution to Photorealistic Image Stylization (Jul. 27, 2018). [cited by applicant]
Liu et al., Unsupervised Image-to-Image Translation Networks (Mar. 2, 2017). [cited by applicant]
Mao et al., Least Squares Generative Adversarial Networks (Feb. 24, 2017). [cited by applicant]
Miyato et al., Cgans With Projection Discriminator (Aug. 15, 2018). [cited by applicant]
Nilsback et al., Automated flower classification over a large number of classes (Dec. 2008). [cited by applicant]
Ronneberger et al., U-Net: Convolutional Networks for Biomedical Image Segmentation (May 18, 2015). [cited by applicant]
Sun et al., Correlation Alignment for Unsupervised Domain Adaptation (Dec. 6, 2016). [cited by applicant]
Sun et al., Deep High-Resolution Representation Learning for Human Pose Estimation (Feb. 25, 2019). [cited by applicant]
Wang et al., High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs (Jun. 2018). [cited by applicant]
Wang et al., ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks (Sep. 17, 2018). [cited by applicant]
Zhu et al., Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks (Mar. 2017). [cited by applicant]
Russian Search Report dated Oct. 23, 2020, issued in Russian Patent Application No. 2020113037. [cited by applicant]
Russian Office Action dated Oct. 26, 2020, issued in Russian Patent Application No. 2020113037. [cited by applicant]
Russian Decision to Grant dated Mar. 3, 2021, issued in Russian Patent Application No. 2020113037. [cited by applicant]
Korean Office Action dated Mar. 18, 2025, issued in a Korean Patent Application No. 10-2022-7008530. [cited by applicant]
Chinese Office Action dated Apr. 18, 2025, issued in a Chinese Patent Application No. 202080079227.3. [cited by applicant]
European Examination Report dated Nov. 27, 2025, issued in European Application No. 20886588.1. [cited by applicant]
Chinese Office Action dated Jan. 12, 2026, issued in Chinese Application No. 202080079227.3. [cited by applicant]
Chinese Office Action dated Mar. 16, 2026, issued in Chinese Application No. 202080079227.3. [cited by applicant]
Hui Li et al., Image Style Transfer Algorithm Under Deep Convolutional Neural Network, Computer Engineering and Applications, 2020. [cited by applicant]
Lechao Cheng et al., Synthesizing the Four Seasons of a Scene From Large-Scale Web Images, Journal of Computer-Aided Design & Computer Graphics vol. 30 No. 5, May 2018. [cited by applicant]