Fisher-Yates Chaotic Shuffling Based Visual Cryptography Scheme for Multimodal Biometric Authentication

In the existing multimodal biometric authentication technique, there is no integrity in the biometric information, and authentication of end users is also not very secured. In this paper, we propose Fisher-Yates Chaotic Shuffling based Visual Cryptography Scheme (FYCS-VCS) for Multimodal Biometric Authentication. The proposed approach enhances the security of multimodal biometric images by protecting the individual shares which supports in maintaining the integrity and confidentiality. Also by the combination of FYCS –VCS Scheme the intruders cannot view and modify shares without knowing the sequential order of the shares. In this technique, initially two or three (finger print, Iris, Face) biometric modalities are considered and fused into a binary image using Histogram of Oriented Gradients (HOG) and Low-Rank Representation (LRR) scheme.Then using (k, n) VSS scheme, n shares of the binary image is created which are encrypted by applying Fisher-Yates Chaotic Shuffling algorithm.. The shares are reconstructed back by applying reverse of Yates-Shuffling algorithm, to generate the fused image. By extracting the HOG features from the fused image, the respective biometric images are obtained. The reconstructed biometric images are authenticated against the original images of the person by an authentication server. The proposed FYCS-VCS multimodal biometric authentication scheme is implemented in MATLAB and is compared against the VSS scheme. Experimental results have shown that FYCS-VCS attain lesser MSE and higher PSNR values when compared to VSS. Proposed approach reduces storage cost and time complexity.