Markerless Motion Capture System for Deformable and Articulated Objects

我们提出了一种新的基于实例的标记自由运动捕获方法。首先，通过基于视觉外壳的方法和细化过程，从每一帧的多个摄像机视图中获取演员身体和骨骼的体数据。然后，检查骨架的结构，我们可以获得属性图。通过大量实例拓扑图的两步图匹配算法，得到人体部位与骨骼部位之间的对应关系。最后，我们通过检测骨骼的曲率来估计关节的位置。实验结果验证了该方法的有效性。

We propose a novel example based marker free motion capture method. First, volume data of actor's body and its skeleton are obtained from multiple camera views at each frame through visual hull based method and thinning process. Then, examining the structure of the skeleton, we can acquire an attributed graph. By using a two-step graph matching algorithm with many example topological graphs, correspondences between body parts and skeleton parts are obtained. Finally, we estimate the position of the joints examining the curvature of the skeleton. The experiment result shows the validity our approach.