State-of-the-art application of artificial neural network in digital watermarking and the way forward

Several high-ranking watermarking schemes using neural networks have been proposed in order to make the watermark stronger to resist attacks.The ability of Artificial Neural Network, ANN to learn, do mapping, classify, and adapt has increased the interest of researcher in application of different types ANN in watermarking.In this paper, ANN based approached have been categorized based on their application to different components of watermarking such as; capacity estimate, watermark embedding, recovery of watermark and error rate detection. We propose a new component of water marking, Secure Region, SR in which, ANN can be used to identify such region within the estimated capacity. Hence an attack-proof watermarking system can be achieved