Neural network, self-organization and object extraction

Ghosh, A., and S.K. Pal, Neural network, self-organization and object extraction, Paitern Recognition Letters l3 (1992) 387-397. Algorithms for object extraction nsing a nenral network are proposed. A single nenron (processor) is assigned here to every pixel for its operalion in order to implement {he concept of self-organized fealnre mapping. Both global and local information have been used as input feature. SI.alisl.ieal criteria for obtaining the optimal output are snggested. Theore\ical proof for lhe convergence of the algorithms is also given. The algorithms are found Lo work well even fOl noisy input.