Transform Domain Watermarking Based on Compressive Sensing and Scrambling

Abstract

ABSTRACT:Conventional sampling follows Nyquist Criterion which states that effective reconstruction of a signal is possibleonly if the signal is sampled at a rate greater than or equal to twice its largest frequency. Compressive Sensing is a new method by which a signal can be reconstructed with fewer samples. Watermarking is the process of hiding data in a carrier signal. Watermarking has been combined with compressive sensing in this paper. To enhance the security of the watermarked image, it is scrambled by random distribution of pixels. PSNR of the watermarked image and the recovered image is compared with respect to embedding ratio. Transform domain is done in this paper as a modification to spatial domain watermarking. Entire image is embedded in transform domain watermarking rather than just four planes which results in the better result.