A Novel Content Based Methodology for a Large Scale Multimodal Biometric System

摘要

Recently, Content Based Image Retrieval (CBIR) system has drawn enormous attention of researchers because of its efficiency in recognizing images from large databases as well as growing demand from real world applications. According to many, biometrics recognition is one of the most potential applications of CBIR. However, no research work has been published up to date on content based multimodal biometric systems. In this proposal, a content based multimodal biometric system, where color, texture, and shape features are combined to enhance the recognition accuracy of the system, is proposed. The preliminary result of the proposed content based feature fusion method for face recognition demonstrates its potential to boost up the recognition performance of a large scale multimodal biometric system.