Most motion-based tracking algorithms assume that objects undergo rigid motion, which is most likely disobeyed in real world. In this paper, we present a novel motion-based tracking framework which makes no such assumptions. Object is represented by

均值便宜跟踪的示例，是学习PCA特征提取的很好的学习资料，实现六自由度运动学逆解算法。- Example tracking mean cheap, Is a good learning materials to learn PCA feature extraction, Six degrees of freedom to achieve inverse kinematics algorithm.