Biometric Template Spoofing Detection Using Sparse Watermarking Scheme

In this paper, a semi fragile watermarking scheme based on compressive sensing theory proposed for detection of spoofing or modification of biometric template. This proposed scheme is explored sparseness, incoherent and computation security provided by compressive sensing for biometric template protection. The Discrete Cosine Transform (DCT) coefficients of biometric template are used to compute sparse measurement vector which is used as secure watermark information. This sparse measurement vector of biometric template as a watermark is embedding into two LSB of selected block of host biometric template of same individual. Extracted sparse measurement vector is required for CS theory recovery algorithm to reconstruct the original watermark biometric image at detector side. The biometric spoofing can be detected by comparing extracted sparse measurement vector and original sparse measurement vector of watermark biometric.