Statistical Image Watermarking In DWT with Capacity Improvement

Abundant techniques has been widely used to design robust image watermarking schemes but in most cases due significance is not given on capacity and data imperceptibility aspects. Robustness of an image-watermarking scheme is the ability to detect the watermark after intentional attacks and normal audio/visual processes. This paper proposes a well- organized blind watermark detection scheme using DWT coefficients. Discrete Wavelet Transform (DWT) is widely applied to image watermarking applications because it decomposes a cover image into spatial domain as well as frequency domain simultaneously. The proposed method improves the capacity of image watermarking. The proposed paper concentrates on some of the main attributes necessary for image watermarking. They are embedding scheme, maximum likelihood detection, decision threshold, and the Laplacian model for image watermarking. The embedding method is multiplicative and done at second level of DWT decomposition by most favorable choice of the embedding strength. The watermark detection is based on the maximum likelihood ratio. Neyman-Pearson criterion is used to reduce the missed detection probability subject to a fixed false alarm probability. The DWT coefficients are assumed to be modeled using the Laplacian distribution. The proposed method is tested for imperceptibility, robustness, and capacity and proved to have better robustness and better imperceptibility and better capacity than other conventional watermarking techniques that were proposed earlier in literature.