• I am currently working as a Postdoctoral Researcher in the VISICS lab, which is part of ESAT-PSI at Katholieke Universiteit Leuven (KU Leuven), under the mentorship of Prof. Tinne Tuytelaars. I obtained my Ph.D. degree from the VDI Center at the School of Information Science and Technology, ShanghaiTech University, where I was supervised by Prof. Jingyi Yu.

• I earned my Doctorate (Ph.D.) from the University of the Chinese Academy of Sciences, with ShanghaiTech University as my main training institution, in 2021.

• My research focuses on computer vision and 3D representation, including novel view synthesis, volumetric videos, 3D reconstruction, and other 3D-related tasks.

Projects

BG-Triangle: Bézier Gaussian Triangle for 3D Vectorization and Rendering

BG-Triangle can transform multi-view images into vectorized 3D scenes, enabling the rendering of sharp novel views.

CVPR 2025

V^3: Viewing Volumetric Videos on Mobiles via Streamable 2D Dynamic Gaussians

This work presents a novel approach that enables high-quality mobile rendering through the streaming of dynamic Gaussians.

SIGGRAPH Asia 2024 (TOG)

Navigating the Nuances: A Fine-grained Evaluation of Vision-Language Navigation

This study presents a novel evaluation framework for the Vision-Language Navigation (VLN) task. It aims to diagnose current models for various instruction categories at a finer-grained level.

EMNLP 2024 Findings

TeTriRF: Temporal Tri-Plane Radiance Fields for Efficient Free-Viewpoint Video

TeTriRF significantly reduces the storage size for Free-Viewpoint Video (FVV) while maintaining low-cost generation and rendering.

CVPR 2024

VideoRF: Rendering Dynamic Radiance Fields as 2D Feature Video Streams

VideoRF treats dynamic radiance fields as 2D feature streams, leveraging hardware video codecs and shaders for smooth, high-quality rendering on various devices.

CVPR 2024

NeVRF: Neural Video-based Radiance Fields for Long-duration Sequences

NeVRF marries neural radiance field with image-based rendering to support photo-realistic novel view synthesis on long-duration dynamic scenes.

3DV 2024

Neural Residual Radiance Fields for Streamably Free-Viewpoint Videos

ReRF enables highly compressible and streamable radiance field modeling. Our ReRF-based codec scheme and streaming player gives users a rich interactive experience.

CVPR 2023

Multi-view Neural Human Rendering

NHR creates photorealistic free-viewpoint videos (FVV) using multi-view dynamic human captures, utilizing point-based neural rendering techniques.

CVPR 2020