A Novel Image Compression Algorithm Based on Wavelet Transform and Neural Network Prediction

摘要

Image compression is an important issue for storing and sending images in telecommunication systems. In this paper, we propose a lossy image compression method based on using wavelet transform to decompose the image into approximation and details subbands and applying a neural network to predict high level details subbands using low level details subbands. We also use the details subbands in different low levels of image wavelet decomposition for training the neural networks. Using this proposed compression scheme, we have achieved close performance in image qualities for the decompressed images in higher compression ratios compared to other well-known compression methods.