Myungsub Choi

Myungsub Choi

Visiting Researcher

Google Research

About Me

I am currently a visiting researcher at Google. I did my PhD at SNU computer vision lab, advised by prof. Kyoung Mu Lee. I am interested in various topics in computer vision and machine learning, including low-level vision problems (video frame interpolation, image/video super-resolution), attention models in neural networks, meta-learning, etc.

Interests

  • Computer Vision
  • Machine Learning
  • Video Analysis
  • Meta Learning

Education

  • PhD in Electrical & Computer Engineering, 2013~2021

    Seoul National University

  • BE in Electrical & Computer Engineering, 2009~2013

    Seoul National University

Timeline

 
 
 
 
 

Visiting Faculty Researcher

Google Research

Apr 2021 – Present Seoul, Korea (working remotely)
 
 
 
 
 

Intern

Snap Inc.

Jun 2018 – Aug 2018 Venice, California, USA
Summer research internship at Snap Research.
 
 
 
 
 

PhD (MSc & PhD unified course)

Computer Vision Lab, Seoul Nat’l University

Mar 2013 – Feb 2021 Seoul, Korea
 
 
 
 
 

Bachelor of Eng.

Seoul Nat’l University

Mar 2009 – Feb 2013 Seoul, Korea

Projects

Meta-VFI

Adaptive video frame interpolation via meta-learning

CAIN

Using channel attention for video frame interpolation

Awesome-RNN

[Deprecated] Curated list of RNN-related works

Recent Publications

Quickly discover relevant content by filtering publications.

[WACV`21] DynaVSR: Dynamic Adaptive Blind Video Super-Resolution

Most conventional supervised super-resolution (SR) algorithms assume that low-resolution (LR) data is obtained by downscaling …
Suyoung Lee*, Myungsub Choi*, Kyoung Mu Lee
PDF Code Project

[NeurIPS`20] Meta-Learning with Adaptive Hyperparameters

Despite its popularity, several recent works question the effectiveness of MAML when test tasks are different from training tasks, thus …
Sungyong Baik, Myungsub Choi, Janghoon Choi, Heewon Kim, Kyoung Mu Lee
PDF Code

[CVPR`20] Scene-Adaptive Video Frame Interpolation via Meta-Learning

We show that input-aware parameter adaptation via meta-learning can greatly improve state-of-the-art video frame interpolation models.
Myungsub Choi, Janghoon Choi, Sungyong Baik, Tae Hyun Kim, Kyoung Mu Lee
PDF Code Project

[AAAI`20] Channel Attention Is All You Need for Video Frame Interpolation

Prevailing video frame interpolation techniques rely heavily on optical flow estimation and require additional model complexity and …
Myungsub Choi, Heewon Kim, Bohyung Han, Ning Xu, Kyoung Mu Lee
PDF Code Project

[ECCV`18] Task-Aware Image Downscaling

Image downscaling is one of the most classical problems in computer vision that aims to preserve the visual appearance of the original …
Heewon Kim, Myungsub Choi, Bee Lim, Kyoung Mu Lee
PDF Project Poster

Contact

  • 133-508, 1 Gwanak-ro, Gwanak-gu, Seoul 08826