Relaxation with model switching and its application to shape matching

This paper introduces a novel approach for contour-based shape matching named as constructive relaxation matching (CRM). In image matching relying on a particular modeling method, apparently similar images can be judged as being quite distant, according to the nature of the modeling process. In the proposed CRM, the modeling stage for a novel input image contour, commonly done in the same procedure used for modeling the templates, will be included in the procedure of iterative relaxation matching. Upon dynamically constructing the model during relaxation, pairs of contour objects having similar template label assignment probabilities will be unified to make one object. After describing the CRM procedures, the method is applied to simple shape matching problems demonstrating the ability to adaptively model the input image during relaxation. It is shown that the proposed CRM improves the object-label correspondence for evaluation of the image similarities in the following stages of shape matching applications.