Neural network, self-organization and object extraction

Abstract

Algorithms for object extraction using a neural network are proposed. A single neuron (processor) is assigned here to every pixel for its operation in order to implement the concept of self-organized feature mapping. Both global and local information have been used as input feature. Statistical criteria for obtaining the optimal output are suggested. Theoretical proof for the convergence of the algorithms is also given. The algorithms are found to work well even for noisy input.