关键词: 形状上下文, 尺度立方体, 样条插值, 最优尺度, 形状配准

Abstract:

In order to more effectively deal with scale changes of shapes, the scale shape descriptors (SSD) are designed by utilizing shape contexts (SC) or Gaussian shape contexts (GSC) with different scales as basis shape descriptors. Firstly, the scales of shape descriptors as intra-layer scales are determined based on the reference shape. Different inter-layer scales used to establish scale cubes of the test shape. Secondly, we calculate the mean distance between every layer of scale cubes and the test shape, and search the nearest scale of the master shape based on the mean distance. Thirdly, the spline interpolation is used to more accurately estimate the optimal scale. Then shape descriptors are built based on the optimal scale and the point to point correspondences are exactly determined on the basis of which the registration transform is calculated. Experimental results show that the SSC can well overcome the difficulties caused by scale changes in shape registration.

Key words: shape context, scale cube, spline interpolation, optimal scale, shape registration