Digital Image Compression Enhancement Using Bipolar Backpropagation Neural Networks

AbstractIt is well known that the classic image compression techniques such as JPEGand MPEG have serious limitations at high compression rate, the decompressedimage gets really fuzzy or indistinguishable. To overcome this problem, artificialneural networks ANNs techniques are used. In this paper, we propose a bipolarsigmoidal backpropagation BBP algorithm to train a feedforward autoassociativeneural network. The proposed method includes steps to break down large images intosmaller windows for image compression/ decompression processes. A number ofexperiments have been achieved, the results obtained, such as compression ratio andpeak signal to noise ratio PSNR are compared with the performance of linearbackpropagation LBP and standard (sigmoidal) backpropagation SBP schemes