Feature vector extraction and optimisation for multimodal biometrics employing face, ear and gait utilising artificial neural networks

摘要

Cloud computing is the rapidly growing model for providing resources to users over internet. Multimodal biometrics is an upcoming research area to explore for improving the security of cloud. In this work, a novel multimodal biometric fusion system using three different biometric modalities including face, ear, and gait, based on speed-up-robust-feature (SURF) descriptor along with genetic algorithm (GA) is anticipated. Artificial neural network (ANN) is utilised as a classifier for each biometric modality. Our novel approach has been effectively tested by means of dissimilar images analogous to subjects from three databases namely AMI Ear Database, Georgia Tech Face Database and CASIA Gait Database. Before going for the fusion, the SURF features are optimised using GA and cross validated using ANN. It is observed that, the amalgamation of face, ear and gait gives better performance in terms of accuracy, precision, recall and F_(measure).