We propose an algorithm for motion extraction from Time-Varying Mesh (TVM). TVM is a sequence of 3D models made for real-world dynamic objects. TVM store detailed information about 3D objects such as their shape, color, and motion. Therefore, TVM is an attractive technology for next-generation multimedia. However, the management of a TVM database, which is used for archiving and retrieving, is a significant problem because it is difficult to locate and track feature points in TVM. For efficient and effective archive and retrieval systems, motion extraction and tracking is essential. In our method, skeletons are extracted from TVM using a Reeb graph, and motion tracking is achieved by fitting a reference skeleton to the other skeletons. We developed a geodesic function using principal component analysis (PCA) for fast and noiseless skeleton extraction. Robust fitting was achieved by implementing an end-node tracking strategy. In addition, we have developed an efficient motion retrieval system compatible with conventional motion capture data using the extracted skeletons. As a result, we have developed a low-cost query generation technique using motion-capture systems and cross search between TVM and motion capture data.
