WITHDRAWN: An invisible based watermarking technique for biometric image authentication

Abstract

Biometric image authentication is an important and challenging area of research. It involves an analysis of the candidate’s biometrics for security purposes. Exposure to such sensitive information can be a threat to an individual’s identity and system security. In this paper, we have proposed an invisible watermarking scheme to protect biometric information from harmful attacks. Initially, the technique segments the host image into blocks using edge entropy and then selects the appropriate band coefficients using 2D-DWT transformation. The adaptive histogram technique is used to embed the watermarking image properly. Finally, to retrieve the original watermark, a reversible data hiding approach (RDH) and visual cryptography technique (VC) is used that provides safety and security to biometric images. The proposed method was tested on two datasets, namely, CASIA-v5 and UBIRIS-v1 and is evaluated on various measuring metrics such as PSNR, NC, and BER. Further, the proposed technique is analyzed and compared with different traditionally implemented watermarking techniques based on PSNR and NC values. Results show that the presented method outperforms the other methods by achieving best in class results and highest PSNR (62.52, 61.25) and NC values (0.95, 0.94).