Brain-Inspired Dual-Stream Fusion for EEG-to-Image Retrieval
Changmin Lee, Jun-Mo Kim, WooHyeok Choi, Yebin Choi (Advisor: Tae-Eui Kam)
Published Proceedings of the 2025 KOSOMBE Fall Conference, p. 374 (Poster-025)
Proposed a dual-stream EEG encoder inspired by the brain's ventral ("what") and dorsal ("where") pathways. Demonstrated significant accuracy improvements in zero-shot image retrieval.
View Abstract (PDF)
Define--Align--Fuse: Dual-Stream EEG-Vision Alignment using Hierarchical CLIP Representation
Chang-Min Lee, Jun-Mo Kim, WooHyeok Choi, Yebin Choi (Advisor: Tae-Eui Kam)
In Prep Targeting IEEE International Winter Conference on Brain-Computer Interface (IEEE BCI) 2026
Extended research of the KOSOMBE abstract, focusing on detailed ablation studies of fusion mechanisms and cross-subject generalization performance.
Text2VR: A Modular Pipeline for Interactive VR Scenes from a Single Text Prompt
Chang-Min Lee*, Sooyoun Myoung, Moonsik An, Youngin Jin, Eui Chul Lee
In Prep Journal of Korea Game Society
Extended journal version of the Text2VR research, focusing on the modular architecture and user experience evaluation of the generated VR scenes.
Emotion Recognition from rPPG via Physiologically Inspired Temporal Encoding and Attention-Based Curriculum Learning
Changmin Lee, H. Lee, M. Whang
Published Sensors, 25(13), 3995, 2025
Paper Link