Securing Person Identification by Combining Hand Biometric Modalities

Recently, extensive studies have been devoted to hand based biometric recognition systems. Thus, based on hand traits captured under visible light, these systems have attained stimulating results but not so much in terms of high accuracy and robustness against fraud. In order to improve these systems, researchers have incorporated the multi-spectral imaging for capturing the hand images. So, the vein images, of palms and fingers, have excellent properties because they are not altered by factors such as dry or wet skin as well as the data exists inside the body are difficult to forge and they cannot be changed or stolen. Furthermore, hand's images can be captured by a device that can provide at the same time a visible and near-infrared frame and then fusing them for the purpose of improving the accuracy and robustness of biometric system. In this context, a multimodal biometric system, based on hand's images captured in different spectrum bands, is proposed. Unlike many biometric systems in the literature, the experimental results show that the proposed scheme provide an excellent accuracy.