Xiangyang Xue

Xiangyang Xue is a Professor at Fudan University in Shanghai, China, with an extensive research portfolio spanning computer vision, machine learning, and artificial intelligence. His work demonstrates significant contributions to object-centric representation learning, 3D reconstruction, person re-identification, and semantic segmentation. With over two decades of publication history from 1999 to present, he maintains an active research program with numerous collaborations, particularly with researchers like Yanwei Fu, Bin Li, and Yu-Gang Jiang.

Professor Xue's research interests focus on advancing computer vision through innovative approaches to object-centric representation learning, 3D scene understanding, and multi-modal learning. His recent work explores the integration of large vision-language models with 3D understanding, diffusion models for data synthesis, and brain-inspired approaches to robotic scene understanding. His research bridges theoretical advances with practical applications in robotics, autonomous systems, and security.

Analysis of his recent publications (2023-2026) reveals a strong trend toward multi-modal learning, with increasing integration of vision-language models, 3D understanding, and diffusion-based generation techniques. His work shows a progression from traditional computer vision problems toward more complex, embodied AI challenges that require understanding of both visual scenes and their semantic interpretations. Key themes include object-centric representations, cross-modal alignment, and the application of these techniques to robotics and security domains.

Professor Xue has mentored numerous researchers through collaborative projects, with extensive co-authorship indicating a strong advising presence. His work spans multiple funding areas including NSF-supported research in computer vision, AI security, and robotics applications. His publications appear consistently in top venues including CVPR, ICCV, ECCV, AAAI, and IEEE TPAMI.

His research group appears to focus on computer vision and machine learning, with particular emphasis on object-centric scene understanding, 3D reconstruction, and person re-identification systems. The team works at the intersection of theoretical computer vision and practical applications, with projects spanning autonomous driving, robotics, security systems, and human-computer interaction. Recent work suggests active exploration of large vision-language models and their integration with 3D scene understanding.