Watermarking for images using wavelet domain in Back-Propagation neural network

A digital image watermarking technique based on Back-Propagation neural networks (BPNN) is proposed. The BPN is a type of supervised learning neural networks. It is a very popular in neural networks. Using improved BPNN, the watermark can be embed into Discrete Wavelet Transform(DWT), which can reduce the error and improve the rate of the learning, the trained neural networks can recover the watermark from the watermarked images. The proposed method has good imperceptibility on the watermarked image and superior in terms of Peak Signal to Noise Ratio (PSNR). The present work analyzes the performance of wavelet filters on variety of test images. The test images are of different size and resolution.