DoHun Lee

DoHun Lee

Ph.D. Student
Kim Jaechul Graduate School of AI, KAIST
leedh7 (at) kaist.ac.kr


About Me

I am a Ph.D. student at the Kim Jaechul Graduate School of AI, KAIST, advised by Prof. Jong Chul Ye. I received my M.S. in Artificial Intelligence from KAIST in Feb. 2026, and my B.S. in Computer Science from KAIST in Feb. 2024.

My research so far has focused on generative models, especially video diffusion models: improving temporal consistency without additional training, training video diffusion transformers with self-supervised vision encoders, and equipping video-to-video diffusion models with memory for consistent multi-turn editing (during my internship at Adobe Research). I have also worked on diffusion models for inverse problems, such as compressed sensing MRI.

These days, I am particularly interested in long-term world modeling, out-of-sight dynamics, and robotics.

Research Interests

News

Publications

* denotes equal contribution

  1. GeoFix: cross-view depth reprojection and overlap calibration arXiv
    GeoFix: Geometry Fixing for 4D Consistent Generation
    Yeobin Hong, Hyojun Go, Dohun Lee, Jeongjae Lee, Sunghwan Hong, Federico Tombari, Zhaochong An, Jong Chul Ye, Konrad Schindler, Dominik Narnhofer
    arXiv preprint, 2026
    Project arXiv
  2. Memory-V2V overview ECCV
    Memory-V2V: Memory-Augmented Video-to-Video Diffusion for Consistent Multi-Turn Editing
    Dohun Lee, Chun-Hao Paul Huang, Xuelin Chen, Jong Chul Ye, Duygu Ceylan, Hyeonho Jeong
    European Conference on Computer Vision (ECCV), 2026
    Project arXiv
  3. ContextMRI overview MedIA
    ContextMRI: Enhancing Compressed Sensing MRI through Metadata Conditioning
    Hyungjin Chung*, Dohun Lee*, Zihui Wu, Byung-Hoon Kim, Katherine L. Bouman, Jong Chul Ye
    Medical Image Analysis (MedIA), 2026
    Paper arXiv Code
  4. PCA visualization of video diffusion transformer and vision encoder features arXiv
    Improving Video Diffusion Transformer Training by Multi-Feature Fusion and Alignment from Self-Supervised Vision Encoders
    Dohun Lee, Hyeonho Jeong, Jiwook Kim, Duygu Ceylan, Jong Chul Ye
    arXiv preprint, 2025
    arXiv
  5. VideoGuide: base vs. ours CVPR
    VideoGuide: Improving Video Diffusion Models without Training Through a Teacher's Guide
    Dohun Lee*, Bryan Sangwoo Kim*, Geon Yeong Park, Jong Chul Ye
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025
    Project Paper arXiv Code
  6. MotionPrompt sample CVPR
    Optical-Flow Guided Prompt Optimization for Coherent Video Generation
    Hyelin Nam*, Jaemin Kim*, Dohun Lee, Jong Chul Ye
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025
    Project Paper arXiv
  7. ACDC overview arXiv
    ACDC: Autoregressive Coherent Multimodal Generation using Diffusion Correction
    Hyungjin Chung*, Dohun Lee*, Jong Chul Ye
    arXiv preprint, 2024
    Project arXiv

Education

Honors & Scholarships


Powered by Jekyll and Minimal Light theme.