Recognition of partially occluded objects using neural network based indexing

In this paper, a new neural network based indexing scheme has been proposed for recognition of planar shapes. Local contour segment-based-invariants h ave been used for indexing. Object contours have been obtained using a new algorithm which combines advantages of region growing and edge detection. N eighbourhood constraints have been applied on the results of indexing for c ombining hypotheses generated through the indexing scheme. Composite hypoth eses have been verified using a distance transform based algorithm. Experim ental results, on real images of varying complexity of a reasonably large d atabase of objects have established the robustness of the method. (C) 1999 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.