Fragility and Robustness of Binary-Phase-Only-Filter-Based Fragile/Semifragile Digital Image Watermarking

Abstract

In the fragile/semifragile digital image watermarking for image authentication, fragility or sensitivity of the embedded watermark to malicious attacks is an important problem. On the other hand, robustness to some incidental manipulations, such as JPEG compression, is required for the semifragile watermark. Recently, a binary-phase-only-fllter (BPOF)-based watermarking technique has been proposed for image authentication, where the authors embedded a watermark upon the BPOF of the Fourier spectrum of an image into the corresponding magnitude. Then, they performed image authentication by calculating the correlation between the extracted watermark and the phase information computed from the watermarked image. In this paper, we propose to use the identical ratio (IR) between the extracted watermark and the computed BPOF to detect watermark and measure the degree of authenticity. The IR is simpler, computationally efficient, and yields better watermark-detection performance when compared to the correlation-based calculation. In addition, it is found that the technique works better for fragile watermarking when compared to the semifragile watermarking. Furthermore, to accommodate semifragile watermarking, we propose to use the low-frequency ranges for watermark embedding and detection, which improves the required fragility and robustness. The performance of the improved algorithm is verified via experimental results.