Maximizing strength of digital watermarks using neural networks

Several discrete wavelet transform (DWT) based techniques are used for watermarking digital images. Although these techniques are robust to some attacks, none of them is robust when a different set of parameters is used or some other attacks (such as low pass filtering) are applied. In order to make the watermark stronger and less susceptible to different types of attacks, it is essential to find the maximum amount of watermark before the watermark becomes visible. In this paper, neural networks are used to implement an automated system of creating maximum-strength watermarks