GeoNeXt: Video Generative Models as Geometry Learner
European Conference on Computer Vision (ECCV) 2026
Haosen Yang 杨浩森
I am a Ph.D. student at the Centre for Vision, Speech, and Signal Processing (CVSSP) at the University of Surrey, supervised by Dr. Xiatian Zhu.
I am also a Visiting Student at Imperial College London, hosted by Dr. Jiankang Deng.
My current research focuses on unlocking the rich intelligence contained in video data to advance video MLLMs, video generation, World Action Models (WAMs), and Vision-Language-Action models (VLAs). Ultimately, I aim to enable richer interactions between humans and environments, humans and agents, and agents and environments.
01 News
NEW Received the Chinese Government Award for Outstanding Self-Financed Students Abroad! 🎉
VTONeXt, a project I led, was accepted to SIGGRAPH Asia 2026.
GeoNeXt was accepted to ECCV 2026.
LPM received a CVPR 2025 Highlight.
Started visiting Imperial College London, hosted by Dr. Jiankang Deng.
Two papers were accepted to CVPR 2025.
One paper was accepted to AAAI 2025.
Two papers were accepted to NeurIPS 2024.
Started a research internship at Samsung in Cambridge, UK.
02 Experience
drag to explore →
03 Selected publications view all →
CVPRFAM Diffusion: Frequency and Attention Modulation for High-Resolution Image Generation with Stable Diffusion
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2025
CVPRImproving Gaussian Splatting with Localized Points Management
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2025 Highlight
NeurIPSAV-GS: Learning Material and Geometry Aware Priors for Novel View Acoustic Synthesis
Advances in Neural Information Processing Systems (NeurIPS) 2024
NeurIPSRecognize Any Regions
Advances in Neural Information Processing Systems (NeurIPS) 2024