Computer Vision · Multimodal AI · Spatial Intelligence
Software Engineering undergraduate at Chung-Ang University and Undergraduate Research Assistant at the University of Hong Kong.
I am currently an exchange student at HKU's School of Computing and Data Science, with research interests in computer vision, multimodal perception, spatial intelligence, and context-aware AI systems.
About
I am a Software Engineering undergraduate at Chung-Ang University and currently an exchange student at the University of Hong Kong. My research experience spans computer vision, multimodal sensing, medical imaging, and physiological-signal analysis.
I am particularly interested in computer vision, spatial intelligence, multimodal perception, and situated AI systems that can understand real-world environments and interact intelligently with people.
Experience
Situated Reality Lab (SiR Lab)
Conducting research on situated intelligence and multimodal perception for interactive environments. Currently exploring context-aware AI systems and developing research directions involving visual perception, user context, and intelligent assistance.
IXLab
Developed machine-learning and deep-learning pipelines for electrooculography (EOG)-based gaze understanding. Evaluated CNN and Transformer architectures under participant-independent settings using leave-one-subject-out cross-validation.
Research
Robust visual perception and recognition in challenging real-world environments, including domain shift and open-world settings.
Combining visual, contextual, and sensor information to understand people, objects, and environments in physical space.
AI systems that understand user intent and environmental context to provide adaptive and intelligent support in real-world settings.
Publication
E. Lee, J. Kim, J. Chung, G. Lim, and J. Paik
2026 IEIE Summer Conference · Presented · Second Author
Projects
Co-developed a YOLOv5-based domain-generalization framework for thermal UAV search-and-rescue imagery using Fourier Domain Adaptation and thermal-specific augmentation to improve robustness under unseen environments.
Developed ML and deep-learning pipelines for six-channel EOG signals collected from 20 participants. Compared classical classifiers, CNNs, and Transformer architectures using participant-independent evaluation.
Developed a multi-agent validation workflow for synthetic chest X-ray samples and investigated clinically reliable synthetic-data augmentation for medical image classification.
Designing an Android-based community marketplace connecting users who request and provide local services, with features including messaging, multilingual support, a coin-based transaction system, and an AI-powered FAQ assistant.
Contact
I am interested in research opportunities involving computer vision, multimodal AI, spatial intelligence, situated AI, and interactive intelligent systems.
estherkim04@naver.com