Hyeongyu Kim

Ph.D. Candidate at Yonsei University, Seoul, South Korea.

prof_pic.png

02-2123-7446

C516, 50 Yonsei-ro, Seodaemun-gu

Seoul, 03722

I am a Ph.D. candidate at MAI-LAB, Yonsei University, working on deep learning and AI.

My main research interests include Domain Adaptation, especially Test-Time Adaptation, Deep learning on Fourier Ptychography, and MRI Reconstruction / Super-resolution.

Feel free to reach out!

news

Feb 21, 2026 Our papers AcTTA & CD-Buffer have been accepted at CVPR 2026!
Sep 19, 2025 Our paper Buffer TTA has been accepted at NeurIPS 2025!
Mar 01, 2023 Our paper SDC-UDA has been accepted at CVPR 2023!

selected publications

  1. 2026_ACTTA.PNG
    AcTTA : Rethinking Test-Time Adaptation via Dynamic Activation
    Hyeongyu Kim*, Geonhui Han*, and Dosik Hwang
    * Equal Contribution
    CVPR (Accepted), 2026
    Code
  2. 2025_BUFFER.png
    Buffer layers for Test-Time Adaptation
    Hyeongyu Kim*, Geonhui Han*, and Dosik Hwang
    * Equal Contribution
    NeurIPS, 2025
    Paper Link Code
  3. 2023_DIMIX.PNG
    Dimix: Disentangle-and-mix based domain generalizable medical image segmentation
    Hyeongyu Kim*, Yejee Shin*, and Dosik Hwang
    * Equal Contribution
    MICCAI, 2023
    Paper Link
  4. 2023_SDC_UDA.PNG
    Sdc-uda: Volumetric unsupervised domain adaptation framework for slice-direction continuous cross-modality medical image segmentation
    Hyungseob Shin*, Hyeongyu Kim*, Sewon Kim, and 3 more authors
    * Equal Contribution
    CVPR, 2023
    Paper Link Code