Ph.D. Student in Computer Science, Shanghai Jiao Tong University
I am a Ph.D. student at Shanghai Jiao Tong University, advised by Prof. Bo Zhao.
My research interests center on egocentric vision, human motion, and hand-object interaction, with a broader interest in real-to-sim data acquisition methods for embodied intelligence.
Discussions and collaborations are always welcome; feel free to reach out.
*: joint first author; † Project Lead; ✉ corresponding author(s).
arXiv preprint
A unified end-to-end model for world-space egocentric hand reconstruction that predicts point-wise visibility, contact, and distance in a single pass, without separate detection, camera tracking, or contact stages.
Project Page / arXiv / Code
arXiv preprint
A framework for world-space hand-object interaction reconstruction from egocentric videos, built on spatial intelligence preprocessing, upper-body HOI diffusion priors, and test-time optimization.
Project Page / arXiv / Code
Findings of the Association for Computational Linguistics
A data-centric instruction tuning method that selects diverse IFT data and merges similar samples through prompting, achieving stronger fine-tuning performance with a much smaller training set.
IEEE/CVF Conference on Computer Vision and Pattern Recognition
A multimodal image fusion framework that uses Segment Anything priors, bi-level optimization, and downstream task feedback to improve visual quality and semantic perception efficiency.
IEEE International Conference on Acoustics, Speech and Signal Processing
A neural architecture search approach for infrared-visible image fusion, designed to balance upstream fusion quality and downstream semantic segmentation performance.
AAAI Conference on Artificial Intelligence
A dual-search framework that combines neural architecture search and loss-function search for multi-exposure image fusion, improving visual quality through automatic optimization.
Shanghai Jiao Tong University. Advisor: Prof. Bo Zhao.