IP Library › Granted Patent US 12,210,587
Granted Patent B2
US 12,210,587 · App. 17/512,312 · Granted Jan 28, 2025

Unsupervised super-resolution training data construction

Inventors: Aleksai Levinshtein (Ontario, CA); Xinyu Sun (Toronto, CA); Haicheng Wang (Toronto, CA); Vineeth Subrahmanya Bhaskara (Toronto, CA); Stavros Tsogkas (Toronto, CA); Allan Jepson (Toronto, CA)
Assignee: SAMSUNG ELECTRONICS CO., LTD.
G06F18/2148G06N3/045G06N3/088G06T3/4053G06T5/20G06T2207/20081
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 12,210,587
App. No.
17/512,312
Granted
Jan 28, 2025
Kind
B2
Abstract

A method for training a super-resolution network may include obtaining a low resolution image; generating, using a first machine learning model, a first high resolution image based on the low resolution image; generating, using a second machine learning model, a second high resolution image based on the first high resolution image and an unpaired dataset of high resolution images; obtaining a training data set using the low resolution image and the second high resolution image; and training the super-resolution network using the training data set.

Claims (23)

1. A method for training a super-resolution network, the method comprising:

obtaining a low resolution image;

generating, using a first machine learning model, a first high resolution image based on the low resolution image;

generating, using a second machine learning model, a second high resolution image based on the first high resolution image and an unpaired dataset of high resolution images, by minimizing an adversarial loss, a cycle loss, and a low-frequency content preservation loss that is defined as a distance between an input and an output of the second machine learning model after passing the input and the output through a low-pass filter;

obtaining a training data set using the low resolution image and the second high resolution image; and

training the super-resolution network using the training data set.

2. The method of claim 1 , wherein the generating the first high resolution image comprises generating the first high resolution image by minimizing an image loss function and a kernel loss function.

3. The method of claim 1 , wherein the generating the first high resolution image comprises generating the first high resolution image by identifying a blur kernel.

4. The method of claim 1 , wherein the generating the second high resolution image comprises generating the second high resolution image by training a generator that performs domain adaptation between the first high resolution image generated by the first machine learning model and a real high resolution image.

5. The method of claim 1 , wherein the training the super-resolution network comprises training the super-resolution network to obtain an input low resolution image captured by a user device, and output a high resolution image.

6. A device for training a super-resolution network, the device comprising:

a memory configured to store instructions; and

a processor configured to execute the instructions to:

obtain a low resolution image;

generate, using a first machine learning model, a first high resolution image based on the low resolution image;

generate, using a second machine learning model, a second high resolution image based on the first high resolution image and an unpaired dataset of high resolution images, by minimizing an adversarial loss, a cycle loss, and a low-frequency content preservation loss that is defined as a distance between an input and an output of the second machine learning model after passing the input and the output through a low-pass filter;

obtain a training data set using the low resolution image and the second high resolution image; and

train the super-resolution network using the training data set.

7. The device of claim 6 , wherein the processor is configured to generate the first high resolution image by minimizing an image loss function and a kernel loss function.

8. The device of claim 6 , wherein the processor is configured to generate the first high resolution image by identifying a blur kernel.

9. The device of claim 6 , wherein the processor is configured to generate the second high resolution image by training a generator that performs domain adaptation between the first high resolution image generated by the first machine learning model and a real high resolution image.

10. The device of claim 6 , wherein the first machine learning model is a deep image prior.

11. The device of claim 6 , wherein the second machine learning model is a generative adversarial network (GAN).

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Oct 27, 2021
From: LEVINSHTEIN, ALEKSAI; SUN, XINYU; WANG, HAICHENG; BHASKARA, VINEETH SUBRAHMANYA; TSOGKAS, STAVROS; JEPSON, ALLAN
To: SAMSUNG ELECTRONICS CO., LTD.
Reel/Frame 057937/0306 →
Continuity (3)
Provisional Application 63113368 · Nov 13, 2020
Provisional Application 63107801 · Oct 30, 2020
Related Publication 20220138500A1 · May 5, 2022
References Cited (28)
US 9349164B2 · Wang et al. · 2016 [cited by applicant]
US 9836820B2 · Tuzel et al. · 2017 [cited by applicant]
US 11232541B2 · Wang et al. · 2022 [cited by applicant]
US 11373274B1 · Yoon · 2022 [cited by examiner]
US 11763544B2 · Zhao · 2023 [cited by examiner]
US 11790502B2 · Chen · 2023 [cited by examiner]
US 20180068430A1 · Sang et al. · 2018 [cited by applicant]
US 20180075581A1 · Shi · 2018 [cited by examiner]
US 20190333219A1 · Xu · 2019 [cited by examiner]
US 20200111194A1 · Wang et al. · 2020 [cited by applicant]
US 20210192286A1 · Liou · 2021 [cited by examiner]
US 20220005157A1 · Shu et al. · 2022 [cited by applicant]
US 20220138500A1 · Levinshtein · 2022 [cited by examiner]
US 20230146181A1 · Meshkin · 2023 [cited by examiner]
US 20230214664A1 · Kudo · 2023 [cited by examiner]
JP 6611107B · 2019 [cited by applicant]
KR 1020210090159A · 2021 [cited by applicant]
KR 1020210116922A · 2021 [cited by applicant]
WO 2019220095A1 · 2019 [cited by applicant]
Y. Yuan, S. Liu, J. Zhang, Y. Zhang, C. Dong and L. Lin, “Unsupervised Image Super-Resolution Using Cycle-in-Cycle Generative Adversarial Networks,” 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Wo… [cited by examiner]
Reda, Fitsum A., et al. “Sdc-net: Video prediction using spatially-displaced convolution.” Proceedings of the European conference on computer vision (ECCV). 2018. (Year: 2018). [cited by examiner]
Lee, S., et al., “Generation and Evaluation of Unpaired Near-Infrared Face Images Using CycleGAN”, Journal of Digital Contents Society, vol. 21, No. 3, Mar. 2020, pp. 593-600. [cited by applicant]
Wang et al., “ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks”, 2018, ECCV, 16 pages total. [cited by applicant]
Ji, et al., “Real-World Super-Resolution via Kernel Estimation and Noise Injection”, 2020, CVPRW, 10 pages total. [cited by applicant]
Zhou, ett al., “Kernel Modeling Super-Resolution on Real Low-Resolution Images”, 2019, IEEE/ICCV, 12 pages total. [cited by applicant]
Ulyanov, et al., “Deep Image Prior”, 2017, Cornell University, 25 pages total. [cited by applicant]
Search Report (PCT/ISA/210) issued Feb. 9, 2022 by the International Searching Authority in International Patent Application No. PCT/KR2021/015450. [cited by applicant]
Written Opinion (PCT/ISA/237) issued Feb. 9, 2022 by the International Searching Authority in International Patent Application No. PCT/KR2021/015450. [cited by applicant]
Cited By (1)
US 12,670,744