Sparse Watermarking Technique for Improving Security of Biometric System

Any biometric system is vulnerable against various attacks. One of the attacks is on the communication channel between the system database and matcher module of the biometric system. Therefore, this paper proposed a sparse watermarking technique based on compressive sensing (CS) theory framework for security against this kind of attack. In this proposed sparse watermarking technique, fingerprint image is taken as watermark information. This fingerprint image is converted into sparse measurements using compressive sensing (CS) theory and embeds these sparse measurements into wavelet coefficients of standard image to generate a watermarked image. The structural similarity measure index (SSIM) quality measure is used for authentication between original watermark fingerprint image and reconstructed watermark fingerprint image. The experimental results show that proposed sparse watermarking technique does not affect verification and authentication performance of a biometric system. The proposed sparse watermarking technique has a good performance compared to the existing watermarking technique in the literature.