Junhui Lin

Hi! I'm a first-year M.S. student in Computer Science at the University of Maryland, College Park. Before joining UMD, I studied Artificial Intelligence at Beijing Normal University in Beijing, China, where I developed a broad interest in computer vision, multimodal learning, and generative models.

I want to thank my mentors Yuanfei Huang, Bonan Li, Ziwen Liu, Yanlin Luo, and my friends like Shujun Yao, Jiale Cui.

My current interests lie in vision/multimodal learning, image/video generation, and robot learning. In the long term, I hope to develop intelligent systems that can efficiently learn from large-scale visual data, reason across modalities, and interact with complex real-world environments.

A quick personal plug: I’m very familiar with the U.S. CS/AI/Robotics master’s application process. During my 2026 application cycle, I was fortunate to receive offers from UMD MSCS, CMU MSCV, CMU MSAIE, UPenn MSR, and UMich MSR. If you have any questions about applying, feel free to email me — happy to help!

CV / Research Statement / Email / Scholar / Github

Research

(* Equal Contribution; ‡ Corresponding Author)

Undergraduate Research

	DARE: Do LLMs have Detail-Awareness in Research Elucidation Jiale Cui, Kaixi Zhong, Tianran Shen, Junhui Lin‡, Tianhao Liu‡ In collaboration with Zhejiang University Under review, 2026 [PDF]
	BlazeVGGT: 13× Faster VGGT at 2,000 Frames via Camera Token without Training Shujun Yao, Junhui Lin, Congying Han, Tiande Guo, Ziwen Liu‡, Bonan Li‡ In collaboration with University of Chinese Academy of Sciences Under review, 2026 [PDF]
	UniVid: Parameter-Efficient Post-Training for Unified Video Understanding and Generation Jiabin Luo, Junhui Lin, Zeyu Zhang, Biao Wu, Meng Fang, Ling Chen, Hao Tang‡ In collaboration with Peking University Under Review, 2026 [Website] [PDF] [Code]
	DeflareMamba: Hierarchical Vision Mamba for Contextually Consistent Lens Flare Removal Yihang Huang, Yuanfei Huang‡, Junhui Lin, Hua Huang ACMMM, 2025 [Code] [PDF]
	The Impact of Secondary Task’s Perceived Value on Individuals' Creativity in Divergent Thinking Tasks Yunxuan Xing, Zhuoyi Sha, Junhui Lin, Xingchen Zhou‡ AHFE, 2025 [Paper]

Projects

Undergraduate Projects

	EmoVision: An Intelligent Psychological Healing Platform Based on Multimodal Emotion Recognition Junhui Lin, Huang Yang, Degang Wei, Yachun Fan‡ Open-Sourced(Congratulations we got nearly 600 stars) and won the third in National AI Innocation Competition held by Zhejing University and Baidu Inc., 2025 [Video] [Website] [Code]
	VRTI: Immersive Virtual Physics Laboratory Based on Semi-Physical Tangible Interaction Junhui Lin, Shujun Yao, Wenzhuo Li, Yanlin Luo‡ National College Students' Innovation and Entrepreneurship Training Program — Completed with an Excellent* Rating, 2025*

Awards

China National Scholarship, 2025 — Top 1% (Ministry of Education)

National Third Prize, China Collegiate Computing Contest – Artificial Intelligence Innovation Contest (C4-AI), 2024 — EmoVision, 58/4000 (Award, Project)

Second Prize, National Undergraduate Mathematics Competition, 2023 — Top 12% (Competition)

Academic Service

Reviewer, CVPR 2026 J2A Workshop (J2A: Generative AI for Movie-Grade Video Production Workshop)

Technical Program Committee Member, MMM 2027 (33rd International Conference on Multimedia Modeling)