Search

It is a major area in computer vision to recognize 3 D objects from 2 D images. One of difficulties of object recognition is caused by that 3 D objects are observed in various shapes in 2 D images. The aspect graph approach is one of available methods to solve the problem. This paper describes a new algorithm for creating aspect graphs for general polyhedral objects. In the proposed algorithm, we use a closed space centered at the origin of an object as a viewing space. We divide the viewing space into sets of viewpoints according to differences of aspects. An aspect graph is constructed by performing set operations on the sets of viewpoints. The set operations are executed by substituting solid polyhedra for the sets of viewpoints. This algorithm enables us (1) to classify aspects of a general polyhedral object under perspective projection, (2) to solve the problem of occlusion caused when an object has nonconvex parts, (3) to construct aspect graphs automatically.