Shape description and retrieval are two challenging research topics in computer graphics and vision. It has been commonly acknowledged that the verbal description of a shape does not give good retrieval results for 3D models. A descriptor which captures information about the geometry and/or topology of shape is needed for the use in the retrieval process. A shape descriptor should be compact enough to guarantee efficient human computer interaction., Moreover, it is often required that a shape descriptor be robust to the position of a model in space and to its size, given that 3D models could be oriented randomly and could have different dimensions.
展开▼