Research Overview

#

# Visual Understanding & Recognition We develop AI algorithms for image and video understanding across diverse visual recognition tasks. Our research covers image classification, object detection, segmentation, video understanding, and multimodal visual perception for intelligent vision systems. Keyword: Image Understanding, Video Understanding, Object Detection, Segmentation, Visual Recognition # Reliable & Generalizable Visual AI We develop vision AI that can reliably operate beyond its training environments. Our research investigates out-of-distribution (OOD) detection, domain generalization, and leverages vision foundation models, Vision-Language Models (VLMs), Large Multimodal Models (LMMs), and Mixture-of-Experts (MoE) to improve robustness and generalization. Keyword: Out-of-distribution (OOD) Detection, Domain Generalization, VLM/LMM, Mixture-of-Experts (MoE) # Efficient Learning & Embeeded Vision AI We investigate efficient learning and adaptation techniques for vision and multimodal AI. Our research includes prompt learning, parameter-efficient fine-tuning (PEFT), domain-specific model adaptation, efficient foundation model utilization, and on-device deployment for practical AI systems. Keyword: Prompt Learning, PEFT, Domain Adaptation, On-Device AI, Embedded Vision # Intelligent Vision Systems & Applications We apply advanced computer vision and multimodal AI to solve real-world problems. Our research includes intelligent visual perception, event and anomaly detection, human-centered AI, and practical vision applications across diverse domains. Keyword: Event Detection, Anomaly Detection, Intelligent Vision Systems, AI Applications

Research Projects

#