Jun-Sang Yoo

Computer Vision · Multimodal Learning

Korea University · Seoul

Jun-Sang Yoo

Exploring visual intelligence across images, video, human behavior, and language.

I am an Integrated M.S. & Ph.D. student in the Deep Image Processing Lab at Korea University, advised by Prof. Seung-Won Jung. My research connects low-level image restoration, temporal video modeling, human-centered understanding, and multimodal learning with text and large language models.

View publications Contact
Research arc Image → Video → Human → Language From low-level visual detail to multimodal reasoning
8Peer-reviewed publicationsInternational conferences & IEEE journals
2021—Korea UniversityIntegrated M.S. & Ph.D. program

Research

Across the visual stack

My research connects image-level detail, video dynamics, human-centered understanding, and language-based multimodal learning.

01

Image Restoration

Low-level vision for super-resolution, reference-based learning, efficient quantization, and fine-detail recovery.

02

Video & Motion

Frame interpolation, optical flow, temporal consistency, and motion-aware representation learning.

03

Human-Centered Vision

Human pose and shape, skeleton-based action recognition, and behavior-centered visual understanding.

04

Multimodal Learning

Connecting image, video, and human motion with text-conditioned generation and language-model reasoning.

News

Recent updates

  1. ECCV 2026

    MASS accepted for publication

    Motion-Aligned Selective Scan for flow-based video frame interpolation.

  2. ICIP 2026

    T2M4AR accepted for publication

    Text-to-motion generation for skeleton-based action recognition.

  3. IEEE TIP

    RefQSR accepted

    Reference-based quantization for image super-resolution networks.

  4. IEEE T-ITS

    Human pose and shape estimation paper accepted

    In-vehicle human understanding with the HIVE dataset.

Selected work

Publications

See all publications

Background

Education

Feb 2021 — Present
Graduate

Korea University

Integrated M.S. & Ph.D. in Electrical Engineering

Deep Image Processing Lab · Advisor: Prof. Seung-Won Jung

Mar 2015 — Feb 2021
Undergraduate

Konkuk University

B.S. in Electronic Engineering

Seoul, Republic of Korea

Contact

Research across vision and multimodal learning

For academic inquiries, collaboration, or questions about image, video, human-centered vision, and multimodal learning, email is the most direct channel.

junsang7777@korea.ac.kr