Matching Noisy Outline Contours Using a Descriptor Reduction Approach

Abstract

Shape Matching is an important area in computer vision researches. We propose in this paper a method to match two outline shapes. Assuming that shapes are stored in the database using their textual descriptors, an iterative process is used to reduce descriptors. After the reduction process, the textual descriptors can be compared in order to perform the matching process. The Textual smoothing is done by applying transformations and reductions of the textual descriptors of shapes to be matched.