Contribution to a marker-free system for human motion analysis

摘要: This paper presents a novel approach to human gait analysis using marker-free system. The devised acquisition system is composed of three synchronized and calibrated charge coupled device cameras. aim this work recognize in gray level image sequences the leg walking reconstruct it three-dimensional space. An articulated threedimensional (3D) model body, based on use tapered superquadric curves, first introduced. A motion-based segmentation, morphological operators, then applied order extract boundaries motion. reconstruction process, least median squares regression next performed, determine location body 3D Finally, spatial coherence imposed reconstructed curves better fit anatomy take into account model. Each stage proposed methodology has been tested both synthetic images real world humans. obtained results suggest that quite promising should be useful study gait.