Seunghwan Lee | Sungkyunkwan Univ.

about me

I’m a second-year M.S. student at Sungkyunkwan University, where I am advised by Prof. Sungeun Hong. I am broadly interested in model efficiency and transfer. My current research focuses on multi-task model merging, test-time adaptation, and efficient machine learning.

educations

[Mar. 2024 - present] M.S. Student, Dept. of Immersive Media Engineering, Sungkyunkwan University, Korea
[Mar. 2017 - Feb. 2024] B.E., Dept. of Information and Communication Engineering, Inha University, Korea

publications

Equal contribution are denoted by *

ICCV

Task Vector Quantization for Memory-Efficient Model Merging

Youngeun Kim*, Seunghwan Lee*, Aecheon Jung*, Bogon Ryu, Sungeun Hong

International Conference on Computer Vision (ICCV), 2025.

PDF Code Project Page

Why Train Everything? Tint a Single Layer for Multi-task Model Merging

Aecheon Jung, Seunghwan Lee, Dongyoon Han, Sungeun Hong

preprint

PDF Code

PRL

Prototypical class-wise test-time adaptation

Hojoon Lee, Seunghwan Lee*, Inyoung Jung*, Sungeun Hong

Pattern Recognition Letters, 2025.

PDF

projects

[Sep. 2024 - Aug. 2025] Sensor Fusion and Missing Modality Handling for Occluded Instance Segmentation in Autonomous Driving
M.S. Students Fellowship by National Research Foundation of Korea (NRF)

honors and awards

[May. 2025] 3rd place, Infrared Instance Segmentation Challenge, ICRA 2025 (Hanwha Systems)
[Feb. 2025] Honorable Mention for Outstanding Paper Award, IPIU 2025
[Dec. 2023] Excellence Award, Capstone Design for Information and Communication Engineering, Inha University

internships

[Feb. 2023 - Feb. 2024] Visual Computing and Learning Lab @ Inha Univ.
[Aug. 2020 - Nov. 2020] TG Consulting @ TG