Fuzzy-Neuro based Robust Digital Image Watermarking Technique

Abstract: The major source of communication in present day word is digital media (multimedia or digital images), as digital media can be easily manipulated, protection of data by legitimate means becomes difficult. Watermarking offers a better solution to this problem. Over the years many have proposed better watermarking algorithms for copy right protection, i.e. increasing the robustness of the watermarked data. Artificial intelligence techniques for watermarking have proven to offer better robustness and give a blind watermarking approach. In this paper, a blind and robust image watermarking algorithm is proposed using hybrid intelligence techniques i.e. combination of Artificial Neural Networks and Fuzzy Logic. The proposed method uses Fuzzy-Neuro system to offer combined advantages of both Artificial Neural Networks and Fuzzy Logic the input to the system is given from the Human Visual System (HVS) model and the output thus generated is used in watermark embedding. The proposed system gives better robustness against many image processing attacks like rotation, sharpening, image contrast attacks, etc.

Keywords: Robustness, Artificial Neural Networks, Fuzzy Logic, Fuzzy-Neuro system, Human Visual System.