A NEW SHAPE REPRESENTATION SCHEME AND ITS APPLICATION TO SHAPE-DISCRIMINATION USING A NEURAL-NETWORK

A new method of shape representation and feature extraction is suggest ed. A shape is approximated by a constant-point polygon in which betwe en any two adjacent vertices of the polygon, the number of points on t he contour of the shape is constant. This representation is applicable for both concave and convex shapes and there is no chance of missing any spikes on the boundary. The sequence of the angle of variation bet ween two consecutive line segments is taken as the primary feature (re presentation of the shape). This sequence is then modelled by an autor egressive (AR) process and the least square error estimate of the AR c oefficient vector is used as input to a multilayer perceptron (MLP) ne twork for learning and classification. Robustness of the shape represe ntation scheme and the MLP classifier is also investigated empirically . Adaptive AR modelling is used for estimating the numerically stable and robust coefficient vector.