SKELETONS FROM DOT PATTERNS - A NEURAL-NETWORK APPROACH

Boundary detection is a well studied problem in the context of shape e xtraction from dot patterns and digital images. For images, particular ly binary images, another frequently encountered issue is finding the skeleton of the object. Unfortunately, in the case of dot patterns, th e skeletonization problem has not received much attention due to the l ack of a proper definition of a dot pattern skeleton. We present a met hod, using artificial neural networks, to extract the skeletal shape o f a dot pattern and demonstrate that the skeleton thus obtained is clo se to the perceptual skeleton, The neural network model proposed here is a modified version of Kohonen's self-organizing model. It is dynami c in the sense that processors can be inserted (Or deleted) during the learning process. Unlike in Kohonen's map, the number of processors h ere need not be known a priori, (C) 1997 Published by Elsevier Science B.V.