Shape retrieval based on Haar's function and hierarchical evolution algorithm

A robust shape matching approach should be invariable to rotation, displacement, scale as well as independence to the distortion caused by noise and segmentation error. The outline of a shape image is described using Haar's function. A multi-scale shape matching approach is presented based on discrete curve evolution. Some algorithm about the hierarchical polygon evolution and the comparability measurement degree for two outlines are put forward. The experimental studies are carried using Haar's transformation under different maximum transformation error. The results shows that the approach studied not only decomposes 2D object into polygonal curve but also induces a hierarchical structure of shape, and that combining "coarse level" with "fine level" in hierarchical similar matching, the retrieval of shape image can be effective for both retrieval accuracy and efficiency.