Yin's Homepage‎ > ‎

Human Motion Tracking

Human Motion Tracking via Multiple Cameras

       Tracking of the human body, also called motion capture or posture estimation, is a problem of estimating the parameters of the human body model (such as joint angles) from the video data as the position and configuration of the tracked body change over time. A reliable motion capture system has many potential applications, such as virtual reality, surveillance systems, sign language translation, and gait recognition for identification. Besides, the motion parameters can be used for motion analysis in applications referring to personalized sports training, choreography, or clinical studies of orthopedic patients.
       Currently available commercial systems for motion capture require the subject to wear special markers, body suits or gloves. We present a markerless and automated system for motion capture via camera networks under complex environments. The system includes both the voxel reconstruction and the motion tracking. While most researchers have taken the approach of working directly with the image data, our system reconstruct the 3D voxel for each frame as input to model acquisition and tracking.

 Overview of Our Approach


Initialization Examples(from 3 views)

Tracking and Rendering Results of a Jumping Sequence

      Related Publication

    1. Junchi Yan, Yin Li, Enliang Zheng and Xuncai Liu. An Accelerated Human Motion Tracking System Based on Voxel Reconstruction under Complex Environments, The Asian Conference on Computer Vision (ACCV) 2009