Real-time and Marker-free 3D Motion capture for Home Entertainment Oriented Applications
Résumé
We propose an efficient real-time method for marker-less 3D human motion capture that requires a single computer. Using input from two calibrated webcams, an extended "Shape from Silhouette" algorithm reconstructs the filmed person in real-time. Fasts 3D shape and 3D skin parts analysis provide a robust and real-time system for human full-body tracking. Animation skeleton and simple morphological constraints ease the motion capture process. Thanks to fast and simple algorithms, and appliance to low-cost cameras, our system is well suited for home entertainment device. Results on long video sequences with fast and complex movements, demonstrate our approach robustness.