Face Recognition using Genetic Algorithm and Neural Networks

Abstract

Fingerprints are the most used biometric feature for the person identification. During research it has been observed that there are number of approaches has been implemented for recognition of fingerprints. Different algorithms and techniques have been used for enhancement purpose, minutiae extraction and for matching. In this thesis work a novel technique to enhance our results by using the combination of genetic algorithm and neural network. As we know that both techniques are the world’s best techniques. Genetic algorithm is used for the extraction of minutiae and neural network is used for the recognition of a fingerprint. Firstly, fingerprint image has been enhanced using histogram equalization process so that our algorithm also capable of predicting good results even for dull or low intensity images too. Then Morphological Image Processing operations using MATLAB 2011b has applied for thinning the lines up to the predefined extent so that it will not further destroy or harm our structure of the image. After that we proceed towards predicting of discontinuities using Genetic Algorithm approach based upon nth last segment. Here genetic algorithm is used to find the best possibility for each discontinues segment in an image. Lastly Enhanced image has fed to NN (neural networks) based trained system to diagnose and match finger print with data set. Several features will be used to train the network so that precision will be made in recognition of finger print.