As commercial motion capture systems are widely used, more and more 3D motion libraries become available, reinforcing the demand for efficient indexing and retrieving methods. Usually, the user will only have a sketchy idea of which kind of motion to look for in the motion database. As a result, how to clearly describe the user’s demands is a bottleneck for motion retrieval system. This paper presented a framework that can handle this problem effectively for motion retrieval. This content-based retrieval system supports two kinds of query modes: textual query mode and query-by-example mode. In both query modes, user’s input is translated into scene description language first, which can be processed by the system efficiently. By using various kinds of qualitative features and adaptive segments of motion capture data stream, indexing and retrieval methods are carried out at the segment level rather than at the frame level, making them quite efficient. Some experimental examples are given to demonstrate the effectiveness and efficiency of the proposed algorithms.
A simple and effective image inpainting method is proposed in this paper, which is proved to be suitable for different kinds of target regions with shapes from little scraps to large unseemly objects in a wide range of images. It is an important improvement upon the traditional image inpainting techniques. By introducing a new bijeetive-mapping term into the matching cost function, the artificial repetition problem in the final inpainting image is practically solved. In addition, by adopting an inpainting error map, not only the target pixels are refined gradually during the inpainting process but also the overlapped target patches are combined more seamlessly than previous method. Finally, the inpainting time is dramatically decreased by using a new acceleration method in the matching process.
