Kyeonghun Kim is a Research Assistant at IMSI Research Lab, Seoul National University, specializing in medical imaging, generative AI, and multimodal deep learning applications. He leads the SUDAL Medical Imaging Team and has extensive experience in developing state-of-the-art AI solutions for medical image analysis, including 3D MRI synthesis, diffusion models, and vision-language models.

He is currently pursuing research on conditional abdominal MRI synthesis using segmentation-guided ControlNet pipelines and label-guided diffusion frameworks. His work involves close collaboration with Samsung Medical Center radiologists and NVIDIA researchers, focusing on clinically relevant AI applications in medical imaging.

Research Interests

  • Medical Image Analysis
  • Generative AI & Diffusion Models
  • 3D Medical Imaging
  • Multimodal Deep Learning
  • Computer Vision
  • Foundation Models
  • Hyperspectral Imaging

Education

  • MSc in Artificial Intelligence, 2024

    Kookmin University (GPA: 4.03/4.5)

  • BSc in Computer Science, 2022

    Hankyong National University (GPA: 4.12/4.5)

Recent Publications

This list is updated automatically from the IMSI publication spreadsheet.

  • Loading publications…

View and search all publications →

Professional Experience

  • AI Researcher at GNEWSOFT (Oct. 2024 ~ Present)
  • Medical Imaging Team Lead at OUTTA (Jul. 2023 ~ Present)
  • AI Researcher at Weatherpia (Mar. 2024 ~ Aug. 2024)
  • Data Scientist at LG Uplus (Jan. 2024 ~ Mar. 2024)
  • AI Instructor at SK Telecom (Dec. 2023)
  • AI Software Engineer at AI O2O Inc. (Aug. 2023 ~ Dec. 2023)
  • Data Analyst at OESON (Jan. 2023 ~ Mar. 2023)
  • Software Engineer at NuriFlex Holdings Inc. (Nov. 2021 ~ Dec. 2022)
  • Junior Developer at AST Global (Oct. 2020 ~ Apr. 2021)

Latest

  • July 2026: Presented 3D-LLDM, LoSA-Net, and MMA-Former as oral papers at IEEE ISBI 2026
  • July 2026: Released new arXiv work on medical imaging, PNI prediction, and multimodal learning
  • March 2026: Presented RAPIDS and GRAVEX at NVIDIA GTC 2026
  • February 2026: Presented MAESIL at IEEE ICEIC 2026
  • September 2025: Presented FOSCU at IEEE APCCAS 2025
  • October 2024: Joined GNEWSOFT as an AI Researcher

Certifications & Licenses

Loading credentials…

Key Achievements

  • Built the official IMSI Research Lab website from ground up
  • Developed HyperspectralMAE foundation model achieving SOTA performance
  • Led deep learning bootcamps for 450+ participants (2023) and 650+ participants (2024)
  • Collaboration with NVIDIA researchers on medical AI projects
  • Multiple publications in top-tier medical imaging conferences