Meetei Mayek Unicode Modeling Using Swarm Intelligence and Neural Networks

摘要

The Different techniques have evolved for better optical character recognition for many scripts, yet very little literature has been found for Meetei Mayek script. The current paper exhibits a new approach to model and simulate handwritten Meetei mayek script by using advanced segmentation tools and recognition algorithms. Preprocessing of the acquired images is needed before segmentation and recognition steps; segmentation is done by using PSOFCM segmentation, while multilayer feed forward neural network with back propagation learning is used for the recognition purpose. It may be noted that PSOFCM segmentation proved useful for MRI image processing in our previous paper, the same technique is used for enhancing the characters. The detailed procedures along with the results are discussed in the sections shown below