Digital watermarking in fractional integral transform domain with chaos based random and partially informed detection

This work aims to design and develop a watermarking scheme that is suitable to applications where high imperceptibility is desired along with improved security from malicious attacks, for e.g., the medical images need to be minimally distorted while embedding the information. Unlike other schemes that are majorly based on DWT-SVD or other decomposition methods that have only embedding strength as the secret key, our proposed scheme is based on embedding watermark information in fractional integral transform domain with embedding locations that are sparse and randomly selected with a 2D robust LSCM based chaos. The secret key comprises of transform orders, input parameters of the chaotic map and the watermark strength. Experimental results and evaluations prove that the proposed scheme is able to provide higher security with larger key space and is robust to most of the image processing attacks.