ABSTRACT
Replay attack is a great threat to the biometric system. It incurs almost negligiblecost to tap the channel and play back the intercepted data. This study proposes an optimized SVD based watermarking over iris biometric to mitigate replay attack. The iris image may be divided into two categories – robust regions and non-robust regions(Gupta and Sehgal 2018). In this paper, we apply two-way authentication on iris biometric– SVD based watermarking of non-robust iris regions and user authentication on robust iris regions. This ascertains two-way check on replay attack – firstly, by verification of the watermark and secondly, by user authentication of randomized robust regions. The watermark used here is a combination of system logo as well as system date and time. The locations to embed watermark in components of SVD watermarking are optimally determined. The modified approach for optimized SVD algorithm helps to achieve a high PSNR value of 35.03db and an average NC value of 0.9968, which is important to check the authenticity of watermark. The proposed approach is designed in a manner that it selects afresh subset of robust regions at each authentication. This evades the replay of intercepted data with a system accuracy of 97.85%. The technique has alsovalidatedagainst several image-processing attacks that may arise over the network transmission.
