M.S./Ph.D. Integrated Student
NeuroAI Lab · Kwangwoon University
Developing reliable and interpretable AI for physiological signals,
wearable sensing, and real-world healthcare applications.
I am an M.S./Ph.D. integrated student at the NeuroAI Lab, Kwangwoon University, working at the intersection of biomedical signal processing and artificial intelligence.
My research focuses on extracting meaningful and reliable information from physiological signals, including EEG, ECG, PPG, GSR, and remote PPG.
I am particularly interested in:
- Graph-based representation learning for physiological signals
- Physiological foundation models and transfer learning
- Multimodal and wearable healthcare AI
- Non-contact physiological sensing
- Real-time and personalized health monitoring
|
|
|
|
| Research Project | Description | Main Topics |
|---|---|---|
| NeuroTruth | A multimodal healthcare AI system for real-time craving-state estimation using wearable physiological signals. The system combines biosignal-based prediction, uncertainty assessment, state confirmation, and a safety-aware conversational agent. | Wearable Biosignals · Craving Estimation · Uncertainty-Aware AI · Conversational Agent |
| Video-based Blood Pressure Estimation | Non-contact systolic and diastolic blood pressure estimation from facial videos using remote PPG representations and graph-based modeling of temporal and phase relationships across facial regions. | rPPG · Blood Pressure · Complex-Valued GNN · Inter-ROI Phase |
| Alcohol Craving Biomarker Discovery | Identification and analysis of physiological biomarkers associated with alcohol craving using multimodal EEG, ECG, PPG, and EDA signals acquired during craving-induction experiments. | EEG · ECG · PPG · EDA · Biomarker Analysis |
| Multimodal Affective AI Agent — AI Seoul Tech | Development of an interactive human-centered AI agent that integrates physiological signals, speech, text, and video with foundation models and LLMs to understand users’ emotional and health-related states and provide interpretable, uncertainty-aware responses. | Multimodal AI · Foundation Models · LLM Agent · Explainable AI |
Enhancing EEG-Based Emotion Recognition Using Sparse Dynamic Graph CNN With ℓ₂,₁-Norm
Min Seok Lee, Dae Hyeon Kim, and Young-Seok Choi
IEEE Sensors Journal, Vol. 25, No. 22, 2025
This study introduces structured sparsity into a dynamic graph convolutional network to learn interpretable and efficient connectivity patterns for EEG-based emotion recognition.
Transfer Learning With a Pretrained Foundation Model for Atrial Fibrillation and Flutter Detection on Single-Lead ECG
Min Seok Lee and Young-Seok Choi
The 40th International Technical Conference on Circuits/Systems,
Computers and Communications (ITC-CSCC), 2025
This study investigates transfer learning with a pretrained physiological foundation model for detecting atrial fibrillation and atrial flutter from single-lead ECG recordings across multiple datasets.
A complete publication list is available on my Google Scholar profile.
Developing AI systems that transform complex physiological signals into
meaningful, reliable, and clinically useful information.
My long-term goal is to bridge biosignal processing, artificial intelligence, and real-world healthcare applications through robust, interpretable, and personalized AI.
I aim to develop models that perform well not only in controlled experiments, but also under practical conditions involving individual differences, sensor variability, motion artifacts, and limited labeled data.
NeuroAI Lab
Department of Electronics and Communications Engineering
Kwangwoon University · Seoul, Republic of Korea
Academic collaboration in biomedical AI and biosignal processing is welcome.