Feature Level Fusion of Palm and Face for Secure Recognition

Abstract

Abstract-Biometric user authentication techniques for security and access control have evoked an enormous interest by science, industry and society in the last two decades. But even the best single biometric system suffers from spoof attacks, intra-class variability, noise, susceptibility etc. In the realm of biometrics, the consolidation of evidence presented by multiple biometric sources is an effective way of enhancing the recognition accuracy of an authentication system. This paper proposes an authentication for a multimodal biometric system identification using two traits i.e., face and palmprint at feature extraction level. The training database consists of face and palmprint images. Principal Component Analysis method is used to extract the features from face and palmprints separately. The feature normalization and feature concatenation scheme followed by a dimensionality reduction procedure is adopted to form the feature matrix. The normalized match (distance) scores generated by respective palm and face features before fusion are used to form fused match score. The Euclidean distance and the feature distance are calculated after fusion. All three distances are used to arrive at final decision. Feedback routine implemented between the feature extraction and the matching modules of the biometric system can lead to substantial improvement in multimodal matching performance.

A Single Objective GA and PSO for the Multimodal Palmprint Recognition System

“…Krishneswari and Arumugam prescribed a multimodal method by combining palmprint and fngerprint traits [9]. Bokade and Sapkal reviewed and discussed the feature-level fusion of palmprint and face traits [10]. Bhagat et al developed a multimodal biometric method by joining face and palm vein traits [11].…”

A novel feature biometric fusion approach for iris, speech and signature

“…The average accuracy of fusion classification lies 97-99% where as FAR and FRR lies in between 0.01 and 0.06. The result section has compared the proposed results with other biometric fusion works (Rattani et al, 2007;Bokade & Sapkal, 2012;Nadheen & Poornima, 2013;Dhameliya & Chaudhari, 2013;Veluchamy & Karlmarx, 2016). It is not necessary that each researcher has taken the same biometric sample or same dataset.…”

“…Malode and Sahare (2017) used MFCC for feature extraction and Hidden Markov Model (HMM) for classification in speech recognition system. It has been empirically proven in many publications (Chetty & Wagner, 2005;Chen & Chu, 2006;Rattani et al, 2007;Zhang et al, 2007;Rattani & Tistarelli, 2009;Almayyan et al, 2011;Liau, & Isa, 2011;Bokade & Sapkal, 2012;Park & Kim, 2013;Nadheen & Poornima, 2013;Dhameliya & Chaudhari, 2013;Eskandari et al, 2014;Saleh & Alzoubiady, 2014, Veluchamy & Karlmarx, 2016Haghighat, 2016;Sarhan et al, 2017;Leghari et al, 2018;Carol & Fred, 2018;Supreetha Gowda et al, 2018) that multimodal biometrics systems improve the recognition accuracy by integrating complementary information over unimodal biometrics systems. Features represent rich information about biometrics; fusion at feature level is believed to be bestowing better performance (Veluchamy & Karlmarx, 2016).…”

A multi-instance multi-sample palmprint identification system

“…The information from these sources can be fused at different stages of the biometric recognition system. The fusion can be done before the matching: at sensor-level [5], [6] or at feature-level [7]- [9], or after the matching: at score-level [10]- [12], ranklevel [13]- [15] or at decision-level [16]- [19]. Nevertheless, the use of multiple sensors system raises the cost of the biometric recognition system [4].…”