SangEun Lee

Research Scientist · ETRI, South Korea

Welcome! I'm currently a researcher at ETRI, and I received my M.Sc. in Applied Artificial Intelligence from Sungkyunkwan University, where I was advised by Prof. Eunil Park.

My current focus is on spatial intelligence, multimodal reasoning, and 3D generative AI — building on prior work in visual emotion reasoning with vision-language models.

Always open to discussing research, collaborations, or new ideas — feel free to reach out!

CV Email Google Scholar GitHub LinkedIn

Recent News

Apr 2026

AwardOur team (w/ DXLab) won 2nd place in the VA Prediction Challenge at ABAW (CVPR Workshop 2026).

Apr 2026

PaperAccepted to AI4RWC (CVPR Workshop 2026): Refining Image-to-3D Foundation Models via Geometric Supervision for Industrial Plant Reconstruction.

Nov 2025

PaperAccepted to AAAI 2026: MASP: Multi-Aspect Guided Emotion Reasoning with Soft Prompt Tuning in Vision-Language Models.

Research Interests

Spatial Intelligence

Spatial reasoning and mental modeling from limited observations, enabling AI systems to understand object relationships, scene structure, and navigable 3D environments.

Multimodal Reasoning

Developing vision-language models that reason across images, 3D representations, and language, with applications in spatial understanding, scene interpretation, and human-centric reasoning.

3D Generative AI

Scene-to-3D generation, geometry-aware reconstruction, and adaptation of 3D foundation models for real-world and industrial applications.