Computer Vision
Learning visual representations that remain reliable in complex, changing environments.
Computer Science @ USTC
Building machines that understand 3D worlds.
I’m Jiayue Yang (杨家越), a Computer Science undergraduate at USTC. My current interests sit at the intersection of computer vision, spatial intelligence, multimodal learning, and embodied AI.
在海中也要望月，在谷底也要生花。
About
Researching perception beyond flat images.
I work with Prof. Yan Xia at SPIN-Lab, School of Artificial Intelligence and Data Science, University of Science and Technology of China.
During my research visit at the University of Cambridge, I work under the guidance of Prof. Olaf Wysocki and Prof. Brian Sheil, associated with CV4DT and the Smart Infrastructure Lab.
Research
What I’m exploring
Current emphasis: robust 3D perception and representations that connect geometry, semantics, and action.
Spatial Intelligence
Understanding geometry, structure, and relationships in three-dimensional scenes.
Multimodal Models
Connecting visual, geometric, and language signals for richer scene understanding.
Embodied AI
Building agents that perceive, reason, and act inside physical environments.
Latest news
C-LaV accepted to CVPR 2026
Our work on weather-robust LiDAR place recognition has been accepted to the CVPR 2026 Main Conference.
Journey
Education & experience
2024 — 2028
B.S. in Computer Science and Technology
University of Science and Technology of China
Huaxia Computer Science and Technology Elite Class, Department of Computer Science and Technology.
May 2025 — Present
Research Assistant
SPIN-Lab, USTC
Research in computer vision, spatial intelligence, multimodal large language models, and embodied AI.
Summer 2026
Research Visit
University of Cambridge
Research on 3D computer vision for semantic reconstruction and digital-twin applications.
Publications
Selected work
C-LaV: Conditional Latent Velocity Field Denoising for Weather-Robust LiDAR Place Recognition
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026
Contact