Abstract
The bidirectional self-organizing neural network (BDSONN) architecture when induced with a multilevel sigmoidal (MUSIG) activation function in its primitives, is efficient in segmenting multilevel images. Since, the MUSIG activation function resorts to heuristic class responses, an optimized version of the same, the OptiMUSIG activation function was proposed using genetic algorithm based optimization techniques. As is with most optimization search based techniques, the generation of the class boundaries of the OptiMUSIG activation function is a very time complex procedure. In this article, an attempt has been made to reduce the time complexity of the generation of the optimized class boundaries of the OptiMUSIG activation function using a quantum inspired genetic algorithm (QIGA). Comparative results are demonstrated on three real life gray level images with eight classes with the conventional GA based approach. The QIGA based OptiMUSIG shows significantly better time efficiency than the conventional GA based OptiMUSIG activation function.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call