LAUSR

In this article, a new and improvised image watermarking (IW) technique on colored images is proposed to mitigate robustness and imperceptibility issues found often in the past presented schemes. Most schemes exhibit low robustness due to LSB’s (Least Significant Bit) and MSB’s (Most Significant Bit) based information hiding in the cover image. However, most of these IW schemes have low imperceptibility as the cover image distortion reveals to the attacker due to information hiding in MSB’s. Such IW schemes used to hide the secret message via MSBs and LSBs are susceptible to malicious intruder attacks which results in low robustness, whereas, the attacks on MSB’s have an impact of reduced imperceptibility. Therefore, this paper propounds a digital IW method named as 2D Otsu algorithm, which allows embedding of a grayscale image into a colored host in the wavelet domain. According to this algorithm, the host image is disintegrated into three color bands of blue, red, and green. Each band is divided into small patches leading to the calculation of entropy followed by finding of the threshold that is estimated by taking the average of entropies of all the patches. Wavelet representation of each patch with entropy less than the threshold is given by applying DWT (Discrete Wavelet Transform). Later, by decomposing gray-scale image into binary bits, secret information embedded by quantization technique into optimal wavelet coefficient blocks either encoding wavelet coefficient differences to 0-bits or 1-bits called MultiBit Encoding (MBE). The results obtained show that the imperceptibility and robustness of this proposed model are superior to other proposed watermarking models and this model gives better results than other models at the same payload.