Visual Intelligence LAB at KOREATECH

Research

Research Areas

Medical Image Analysis

We develop artificial intelligence–based methods for medical image analysis to support accurate diagnosis and clinical decision-making. Our research focuses on robust learning frameworks that handle variability across patients, devices, and institutions. By leveraging attention mechanisms, self-supervised learning, and adaptive inference strategies, we aim to enable reliable deployment of AI systems in real-world clinical environments.

Medical Image Analysis
  • Expert-level differentiation of incomplete Kawasaki disease and pneumonia from echocardiography via multiple large receptive attention mechanisms
    • Haeyun Lee, Kyungsu Lee, Moon Hwan Lee, Sewoong Kim, Yongsoon Eun, Lucy Youngmin Eun, Jae Youn Hwang
    • Computers in Biology and Medicine [Paper]
  • Fine-Tuning Network in Federated Learning for Personalized Skin Diagnosis
    • Kyungsu Lee, Haeyun Lee, Thiago Coutinho Cavalcanti, Sewoong Kim, Georges El Fakhri, Dong Hun Lee, Jonghye Woo, Jae Youn Hwang

Image Synthesis

We explore image synthesis techniques tailored for medical imaging to address data scarcity and domain gaps between synthetic and real clinical data. Our work emphasizes realistic degradation modeling and restoration to generate medically meaningful images. These approaches facilitate effective training, benchmarking, and evaluation of medical imaging algorithms under practical clinical conditions.

Image Synthesis
  • Real-Time Self-Supervised Ultrasound Image Enhancement Using Test-Time Adaptation for Sophisticated Rotator Cuff Tear Diagnosis
    • Haeyun Lee, Kyungsu Lee, Jong Pil Yoon, Jihun Kim, Jun-Young Kim
    • IEEE Signal Processing Letters [Paper]

Projects

[ Completed ] 경량형 온디바이스용 다화자 TTS 프로토타입 구현
2025.09 –2025.12
기관 충남연구원
사업 RISE 사업
역할 PI
내용 경량 모델 설계 및 프로토타입 구현(모바일/엣지 환경 가정).
On-device TTS