Reversible Face Watermarking Scheme using Hash Function for Tamper Localization and Recovery

Recently, since digital cameras are widely used, face recognition has attracted great attention. Face recognition has been used in different fields which are airport security management and IC cards for e-commerce. For advanced technologies, biometric data such as face images can be compromised and faked. Hence, security is needed to protect face images in the face recognition system in order to prevent modified and manipulated images. This research proposed an image watermarking scheme for protecting the original face image and maintaining the confidentiality of the information contained in the face image in the image recognition system from unauthorized attacks. The proposed scheme is to detect the tampered sector on the face image after embedding the watermark into the original face images. The proposed scheme able to detect the tampered sector and recover back to its original image. This research also evaluated the recovered face image towards its original image in terms of image quality. This paper also investigated an efficient way to authenticate face image, tamper localization, and lossless recovery for the face image. The imperceptibility of the recovered face image and computational time are evaluated on the watermarked image. The experimental results show that the proposed scheme successfully detects the tampered sector and the scheme able to recover back the tampered face image to its original image. The proposed scheme achieved an average PSNR value of 47.66 dB with a computational time of 1.14 seconds. The proposed scheme has the potential to provide security for authenticating and recovering face images with high imperceptibility and a fast embedding process.