System to improve the Resistance in Data Compression by Evaluating the Performance of Watermarking Algorithm using Multilayer Perceptron

Abstract

ABSTRACT newline newline An advanced watermark is a kind of indication surreptitiously fixed in a scream liberal placard, for example music, visual aid or image information.quotWatermarkingquot is the direction of wrapper up advanced data in a bearer flag need not contain a relationship to the carrier flag. Watermarking is used for protection of content, authentication of content and tamper detection. Watermarking works in 2 stages, first is embedding and second is detecting. In the embedding stage a message and an image, is fed into the embedder algorithm. The algorithm embeds the message within the image resulting in the watermarked image. In the detecting phase, the watermarked image or any image is fed into the detector algorithm, which returns a message representing whether a watermark is present in the image. newline This paper mainly focuses on MLP neural networks to carry out supervised blocks classification in order to get better recital in the detection step. In fact, a supervised learning function based on MLP (Multi-Layer Perceptron) neural network is introduced. The training data consist of a column version of image blocks. As this kind of networks involves supervised learning fix the target output,a real value for each block, a label is associated to it. newline A MLP method consists of one or more hidden layers, neurons in the hidden layers must adjudicate between the key in and out of neural network. Thekey in feature vector is supply into the source nodes in the input layer of the neural network at first. The neurons of the input layer constitute the input signals and apply them to neurons of the first hidden layer. The output signals of the hidden first layer can be used as inputs to then exit hidden layer or the output layer. Finally, the output layer harvest the output results and terminates the neural computing decision. newline The output results of MLP neural network can be used to forecast the presence of the mark with blocks that doesn t belong to the training data. newline newline