Ke Yu

I am currently a 3rd-year PhD candidate in the Multimedia Laboratory, The Chinese University of Hong Kong. My supervisor is Prof. Xiaoou Tang, and I work closely with Prof. Chen Change Loy and Prof. Chao Dong.

I got my B.Eng. degree in the Department of Electronic Engineering, Tsinghua University.

Email  /  Google Scholar  /  GitHub


I'm interested in image restoration, super-resolution and reinforcement learning.

EDVR: Video Restoration with Enhanced Deformable Convolutional Networks

Xintao Wang, Kelvin C.K. Chan, Ke Yu, Chao Dong, Chen Change Loy
CVPR Workshops, 2019
[ PDF ] [ Project Page ] [ Codes]

Our EDVR wins the champions and outperforms the second place by a large margin in all four tracks in the NTIRE19 video restoration and enhancement challenges.

Path-Restore: Learning Network Path Selection for Image Restoration

Ke Yu, Xintao Wang, Chao Dong, Xiaoou Tang, Chen Change Loy
arXiv preprint arXiv:1904.10343
[ PDF ] [ Project Page ] [ Codes ]

We propose Path-Restore, a multi-path CNN with a pathfinder that could dynamically select an appropriate route for each image region.

Deep Network Interpolation for Continuous Imagery Effect Transition

Xintao Wang, Ke Yu, Chao Dong, Xiaoou Tang, Chen Change Loy
CVPR, 2019
[ PDF ] [ Project Page ] [ Codes ]

We propose Deep Network Interpolation (DNI), a simple yet universal approach to attain a smooth control of diverse imagery effects in many low-level vision tasks.

ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks

Xintao Wang, Ke Yu, Shixiang Wu, Jinjin Gu, Yihao Liu, Chao Dong, Yu Qiao, Chen Change Loy
ECCV Workshops, 2018
[ PDF ] [ Project Page ] [ Codes ]

ESRGAN won the champion in the PIRM 2018 Challenge on Perceptual Super-Resolution (3rd track).

Crafting a Toolchain for Image Restoration by Deep Reinforcement Learning

Ke Yu, Chao Dong, Liang Lin, Chen Change Loy
CVPR, 2018 (spotlight)
[ PDF ] [ Project Page ] [ Codes ]

RL-Restore learns a policy to select appropriate tools from a toolbox to progressively restore the quality of a corrupted image.

Recovering Realistic Texture in Image Super-resolution by Deep Spatial Feature Transform

Xintao Wang, Ke Yu, Chao Dong, Chen Change Loy
CVPR, 2018
[ PDF ] [ Project Page ] [ Codes ]

We adopt Spatial Feature Transform (SFT) to generate affine transformation parameters for spatial-wise feature modulation, recovering textures faithful to semantic classes.

Deep Convolution Networks for Compression Artifacts Reduction

Ke Yu, Chao Dong, Chen Change Loy, Xiaoou Tang
arXiv preprint arXiv:1608.02778
[ PDF ] [ Project Page ] [ Codes ]

Fast AR-CNN achieves a speed up of 7.5x with almost no performance loss compared to ARCNN for compression artifacts reduction.

Thanks Jon Barron for sharing the website codes.