Face and palmprint multimodal biometric system based on bit-plane decomposition approach

Abstract

Bit-plane decomposition approach has been introduced \ate\y for single trait biometric system such as face, palmprint, and fingerprint recognition. This approach is able to provide promising high performance rate while reducing the data dimensionality. However, this approach has not been tested on multimodal biometric system which uses more than one biometric trait. Hence, this paper introduces a new multimodal biometric system based on face and palmprint fusion with bit-plane decomposition approach. Pixel level fusion is applied by using simple averaging method before bit-plane feature extraction. Principal Component Analysis is also used on the hybrid face-palm bit planes for further dimension reduction before being classified by Feedforward Backpropagation Neural Network. The experimental results show that the proposed system is able to provide high recognition rate.