Robust Watermarking against Geometric Attacks using Multiple Parameter Discrete Fractional Fourier Transform and Least Significant Bit Technique

Abstract

Watermarking is an important technique used for copyright protection, authentication, tamper detection and hiding secret information in multimedia contents. In this article a new digital watermarking scheme is proposed based on multiple parameter discrete fractional Fourier transform and using least significant bit (LSB) technique. The multiple parameter discrete fractional Fourier transform (MPDFRFT) is the generalization of the discrete fractional Fourier transform (DFRFT) in terms of its different nature of its fractional orders. The MPDFRFT can converge to DFRFT when all of its fractional order parameters having a similar value. The multiple parameter discrete fractional transforms shows it’s superiority over DFRFT because of that it benefitted us with an extra degree of freedom that is provided by its multiple fractional orders. The LSB technique is more preferable because of its lesser effect on the image characteristics. During the implementation of proposed scheme cover image is subdivided and MPDFRFT is applied to each subdivided image to transformed coefficients and watermark images are embedded by using LSB technique. Similarly during watermark extraction process the reverse order of MPDFRFT is applied for reconstruction of original images along with LSB extraction mechanism. The MPDFRFT fractional order parameters are used as secret key to improve the robustness of the system. The robustness of the watermark image is tested under different attacks such as additive noise, cropping, rotation and Gaussian low-pass filtering is analyzed and the results demonstrate that the embedding scheme has good performance of robustness. The imperceptibility of the proposed scheme against the hostile person is also verified by simulations results.