3D Object Retrieval and Recognition

As the development of techniques for modeling, digitizing and visualizing, 3D data are now becoming explosion in number and are widely recognized as the upcoming wave of digital media. In particular, 3D data acquisition techniques make it possible to acquire one 3D object with detailed shape information in a short time. Accordingly, this led to the development of techniques including processing, recognition, categorization, and retrieval for such 3D objects. In this thesis, we studied 3D object retrieval and 3D facial expression recognition within this field. 3D object retrieval is to search for 3D object(s) meeting some specific requirements within a database or the World Wide Web. In this thesis, a novel feature which is called object flexibility, is proposed at a point of a 3D object to describe how the neighborhood of this point is massively connected to the object. This feature is stable to the deformation of objects’ articulations, in addition to commonly concerned linear transforms, i.e., translation, scale, and rotation. A shape descriptor is obtained based on this feature using the bagof-words model. As an application, the descriptor is used to perform 3D object retrieval. Extensive experiments demonstrate its superiority over a variety of existing 3D shape descriptors in the retrieval of articulated objects, as well as its enhancement of other shape descriptors to retrieve generic 3D objects. Facial expression recognition has many applications in multimedia processing, and the development of 3D data acquisition techniques make it possible to identify expressions using 3D shape information. We propose an automatic