A Robust Multi-Biometric System with Compact Code for Iris and Face

Abstract

Multimodal biometric can overcome the limitations of the single biometric trait and gives better classification accuracy. This paper proposes a face-iris multimodal biometric system based on fusion at the matching score level. The iris recognition system is composed of segmentation, normalization, feature encoding and matching. The wavelet transform is used in feature extraction to generate a compact feature vector length of 128 bits; this reduces the computational time and storage of the iris code. The face recognition system is composed of enhancement, feature encoding and matching. The new method called 'Phase-based Gabor Fisher Classifier (PBGFC)' is used in feature extraction; PBGFC employs only 16 Gabor filters, i.e., filters with 2 scales and eight orientations. This fact makes the resulting feature vector for the PBGFC method very compact. The scores from iris and face are then combined using several score normalization and fusion techniques. To validate our approach, experiments are conducted on the iris and face images obtained from the CASIA and ORL datasets respectively. The results show that our multimodal biometric system achieves higher accuracy than both single biometric approaches and the other existing multi-biometric systems based on fusion of iris and face.

A framework for Modified Firefly Algorithm in Multimodal Biometric Authentication System

“…Biometrics is the science of assessing personal characteristics such as the iris, face, fingerprints, retina, palm print, hand geometry, voice, or signatures to secure authentication [1]; [2] and is becoming the technology of the future in the field of security [3]. The biometric system identifies people based on their physiological and/or behavioral features.…”

Dynamic Audio-Visual Biometric Fusion for Person Recognition

“…Fusion at feature level is based on a concatenation of different feature vectors extracted from different biometric modalities to create a new more powerful feature vector with a higher dimensionality that represents the individual more accurately [5,6]. Since the feature sets may contain richer information from biometric modalities than the mere match scores or final decisions, using fusion at the feature level is expected to achieve better recognition results [7].…”