Chapter 12: Contour Rectification and Analysis Using Circular Augmented Rotational Trajectory Algorithm

Abstract

This paper presents a novel circular augmented rotational trajectory (CART) algorithm to compute an R-space based shape descriptors which allow efficient shape matching, generalization and classification. The rotation invariant R-space representation can be used to detect invariant geometric features despite the presence of considerable noise and quantization errors. Moreover, the CART method is corner preserving and can detect the points of discontinuity in a noisy trajectory. Experimental analysis performed on a number of difficult or ambiguous object boundaries show that the CART method can correctly detect and represent the inherent shape and extract their geometric properties. The method's universality, robustness and consistent performance on a variety of difficult shapes make it a power technique for contour representation and analysis.