Wei Xu

Multimodal AI & Human-Centric Systems

Brief Bio.

Hi ๐Ÿ‘‹ I am Wei Xu ( ่ฎธไผŸ in Chinese ), a Ph.D. candidate at USTC advised by Prof. Tingrui Pan and Prof. Kang Li. Before that, I received my masterโ€™s degree in Computer Vision from UESTC in 2021, advised by Prof. Guotai Wang, Prof. Shaoting Zhang, and Prof. Kang Li.

My Research Interests:

  • ๐Ÿง  human-centric multimodal learning
  • ๐Ÿงฉ vision-language models
  • ๐Ÿง human/hand pose estimation
  • ๐Ÿšถ gait analysis and recognition

News

  • ๐Ÿ”ฅ New Project: MedLSAM โ€” Localize and Segment Anything Model for 3D CT Images
    • A complete medical adaptation of Segment Anything Model (SAM) for 3D medical image localization and segmentation. Paper & Code

Projects

๐Ÿงช Resources for human-centric multimodal learning.

ViPLab develops resources and research systems for human-centric multimodal learning, covering pose estimation, hand pose estimation, gait recognition, and gait analysis.

๐Ÿฉบ Foundation model for localizing and segmenting anatomy targets in 3D medical images.

MedLSAM is a localization-and-segmentation foundation model for 3D medical images, designed to segment anatomy targets without additional task-specific annotation.

2D Pose Estimation Pytorch

2D Pose Estimation Pytorch

๐Ÿง PyTorch code for human pose estimation, especially for CAREN-style rehabilitation data.

This project provides an improved framework for accurate 2D and 3D pose estimation in CAREN systems using multi-view videos.

Education

USTC

Ph.D in Computer Vision

SEPT. 2021 - JUL. 2024 (Expected)

University of Science and Technology of China (USTC)

UESTC

Master in Computer Vision

SEPT. 2018 - JUL. 2021

University of Electronic Science and Technology of China (UESTC)

UESTC

Bachelor in Mechanical Engineering

SEPT. 2014 - JUL. 2018

University of Electronic Science and Technology of China (UESTC)

Experience

RA

CV, Biomedical Big Data Center

MAR. 2022 - NOW

BBDC homepage

West China Biomedical Big Data Center, West China Hospital, Sichuan University
West China Hospital-SenseTime Joint Lab

  • Advisor: Prof. Kang Li
  • ๐Ÿšถ Gait Recognition, Human Pose Estimation, Human-centric Multimodal Learning with Vision-language Models

Research Intern

CLD, TacSense

DEC. 2021 - FEB. 2022

TacSense homepage

TacSense Technology (Suzhou) Co., Ltd.

RA

CV, Biomedical Big Data Center

JUL. 2021 - AUG. 2021

BBDC homepage

West China Biomedical Big Data Center, West China Hospital, Sichuan University
West China Hospital-SenseTime Joint Lab

  • Advisor: Prof. Kang Li
  • ๐Ÿšถ Gait Recognition, Gait Analysis

Master Student

HPE, HiLab

MAR. 2018 - JUN. 2021

hilab homepage

Healthcare Intelligence Lab, University of Electronic Science and Technology of China

Bachelor Student

SIM, Digital Design & Simulation Lab

MAR. 2017 - JUN. 2018

SMEE homepage

Digital Design and Simulation Lab, University of Electronic Science and Technology of China

  • Advisor: Prof. Yating Yu
  • โš™๏ธ Simulation and Optimization of High-speed Rotating Experimental Platform for Dynamic Eddy Current Detection

Publications

โœจ * indicates equal contribution. Selected publications are grouped with venue tags, authors, and direct links.

๐Ÿฆต Journal of Biomechanics2025Markerless Kinematics

Evaluation of a smartphone-based markerless system to measure lower-limb kinematics in patients with knee osteoarthritis

Junqing Wang*, Wei Xu*, et al.

๐Ÿงฉ arXiv2023Vision-Language Tuning

ConES: Concept Embedding Search for Parameter Efficient Tuning Large Vision Language Models

Huahui Yi, Ziyuan Qin, Wei Xu, Miaotian Guo, Kun Wang, Shaoting Zhang, Kang Li, Qicheng Lao.

๐Ÿง  arXiv2023Continual Medical AI

Towards General Purpose Medical AI: Continual Learning Medical Foundation Model

Huahui Yi, Ziyuan Qin, Qicheng Lao, Wei Xu, Zekun Jiang, Dequan Wang, Shaoting Zhang, Kang Li.

โœ‹ JVCIR20233D Hand Pose

MTMVC: Semi-supervised 3D hand pose estimation using multi-task and multi-view consistency

Donghai Xiang, Wei Xu, Yuting Zhang, Bei Peng, Guotai Wang, Kang Li.

๐Ÿ”ฌ Clinical Chemistry2022Explainable Medical AI

Expert-level Immunofixation Electrophoresis Image Recognition based on Explainable and Generalizable Deep Learning

Honghua Hu*, Wei Xu*, Ting Jiang, Yuheng Cheng, Xiaoyan Tao, Wenna Liu, Meiling Jian, Kang Li, Guotai Wang.

๐Ÿง  IEEE JBHI2022Brain Segmentation

HMRNet: High-and-Multi- Resolution Network With Bidirectional Feature Calibration for Brain Structure Segmentation in Radiotherapy

Hao Fu, Guotai Wang, Wenhui Lei, Wei Xu, Qianfei Zhao, Shichuan Zhang, Kang Li, Shaoting Zhang.

๐Ÿšถ IEEE T-HMS2022Multiview Pose

Multiview Video-Based 3-D Pose Estimation of Patients in Computer-Assisted Rehabilitation Environment (CAREN)

Wei Xu, Donghai Xiang, Guotai Wang, Ruisong Liao, Ming Shao, Kang Li.

๐Ÿ“ MICCAI2021One-Shot Localization

Contrastive Learning of Relative Position Regression for One-Shot Object Localization in 3D Medical Images

Wenhui Lei*, Wei Xu*, Ran Gu, Hao Fu, Shaoting Zhang, Shichuan Zhang, Guotai Wang.

โš™๏ธ FENDT2018Electromagnetic NDT

Velocity effect in defect detection for ferrite metals by electromagnetic NDT

Fei Yuan, Yating Yu, Na Yu, Wei Xu, Ning Zhao, Guiyun Tian.

Challenges

๐Ÿ… MICCAI 2022 Diabetic Foot Ulcer Challenge (DFUC) 2022

Services

๐Ÿ“ Journal Reviewer

  • IEEE Transactions on Human-Machine Systems (THMS)

๐ŸŽค Conference Reviewer

  • ACM The Web Conference 2024 (ACM TheWebConf 2024)

A Little More

Outside research, I keep a few simple rituals that help me stay curious and reset.

  • ๐Ÿ€ Basketball
  • ๐Ÿšด Riding
  • ๐Ÿ‘Ÿ Sneakers