Robust Perception
Reliable 3D human pose estimation and visual understanding under occlusion and uncertainty.
Computer Vision × AI for Science
Building intelligent systems that remain useful when the world gets messy.
I work across robust perception, medical robotics, and scientific AI—turning mathematical ideas into reliable learning systems for real-world research.
01 / Focus
Research directions
Three connected themes, one goal: dependable AI for complex physical and scientific settings.
Medical Robotics
Frequency-domain visual servoing and intelligent systems for reproducible medical imaging.
AI for Science
Physics-informed agents and learning systems for spectroscopy, materials, and molecular discovery.
02 / Selected work
Research output
PICFormer: Perception-Inference-Consistency Loop for Occluded 3D Pose Estimation
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP, CCF-B)
A closed-loop feedback Transformer architecture for 3D pose estimation under severe occlusion, featuring Visibility-aware Feature Modulation (VFM) and Gated Pyramid Attention (GPA) modules.
Superspreading and Temporal Dynamics of COVID-19 Transmission: Insights from Transmission Settings and Case Detection in Shenzhen
Emerging Infectious Diseases (SCI Q2 Top, Under Review)
A dual-dimension clustering analysis framework for understanding COVID-19 transmission heterogeneity and super-spreading events in Shenzhen.
DCT-Based Robotic-Assisted Photoacoustic Tomography for Vascular Longitudinal and Reproducible Monitoring in Human
IEEE Transactions on Robotics Special Issue on Robot-Assisted Medical Imaging (SCI Q1 Top)
A novel DCT-based frequency domain visual servoing algorithm for robotic-assisted photoacoustic tomography, enabling long-term reproducible vascular monitoring.
03 / Experience
From models to deployed research workflows
2026-03 LLM Algorithm Engineer (Intern) AI for Science Institute (AISI), Beijing 2025-04 AI Algorithm Intern De-Carbon Tech - Southern University of Science and Technology 2024-08 AI Algorithm Intern Xingxin Technology - The Chinese University of Hong Kong
Full experience Collaboration
Interested in robust AI systems for science?
I am open to PhD opportunities, research collaborations, and R&D conversations.
Start a conversation