3D model and motion retrieval: The extended dimensions for web-based learning

Traditionally, web-based learning techniques focus on the extensive use of multimedia data such as images, videos and audios. In this paper, we validate 3D model and motion retrieval as a part of e-learning architecture. New approaches, including a novel shape descriptor for arbitrary 3D polygon models, are proposed as initial solutions to these problems. Preliminary experiments demonstrate the promise of our methods, with potential applications in digital library, interactive learning, etc.