Securing Images by Secret Key Steganography
Younes Benlcouiri, Moulay Chrif Ismaili, Abdelmalek Azizi, Mohammed Benabdellah · 2012
Most watermarking methods introduced that require concealment is less detectable or undetectable, while others exist that are based on the types of concealment that do not require an indiscernibility. It is within this framework and to ensure the security, at storage and transmission of digital data, we propose in this work, a new crypto-watermarking method using symmetrical injection of noise on architecture of multilayer neural networks performed to compression images. Our crypto-watermarking algorithm provides a good compression ratio and a progressive transmission, which reduces the bandwidth. As regards the time necessary for the operations of dissimilation, it may vary depending on the degree of security.