Sangbeom Lim

I am a Ph.D. candidate at Korea University, advised by Seungryong Kim and Paul Hongsuck Seo, and a research intern at Adobe. My research focuses on general perception in video — especially Segment Anything and Matting Anything (VideoMaMa) — and its application to video understanding.

I have interned at Adobe in San Jose, where I was fortunate to work with Seoung Wug Oh and Joon-Young Lee. Earlier, I was a research intern at NAVER Search U.S. working with Dongchan Kim, and a Language AI researcher at the NLP Center of NCSOFT, working with Ayoung Moon and Seungjin Lee.

News

Jun 2026Two papers (SOLA, VIRAL) accepted to ICML 2026 workshops.
Apr 2026Started a research internship at Adobe (San Jose, CA).
Feb 2026VideoMaMa accepted to CVPR 2026.

Publications

* denotes equal contribution. Selected work

2026

VideoMaMa: Mask-Guided Video Matting via Generative Prior

Sangbeom Lim, Seoung Wug Oh, Jiahui Huang, Heeji Yoon, Seungryong Kim, Joon-Young Lee

CVPR, 2026

Work done during my internship at Adobe.

Project Page arXiv Code Demo

Referring Video Object Segmentation via Language-aligned Track Selection

Seongchan Kim*, Woojeong Jin*, Sangbeom Lim*, Heeji Yoon*, Hyunwook Choi, Seungryong Kim

ICML Workshop, 2026

Project Page Paper Code

Visual Representation Alignment for Multimodal Large Language Models

Heeji Yoon*, Jaewoo Jung*, Junwan Kim*, Hyungyu Choi, Heeseong Shin, Sangbeom Lim, Honggyu An, Chaehyun Kim, Jisang Han, Donghyun Kim, Chanho Eom, Sunghwan Hong, Seungryong Kim

ICML Workshop, 2026

Project Page arXiv Code