A Neural Network based Technique for Data Compression

This paper presents a neural network based technique that may be applied to data compression. The proposed technique breaks down large images into smaller windows and eliminates redundant information. Finally, the technique uses a neural network trained by direct solution methods. Conventional techniques such as Huffman coding and the Shannon Fano method are discussed as well as more recent methods for the compression of data and images. Intelligent methods for data compression are reviewed including the use of Backpropagation and Kohonen neural networks. The proposed technique has been implemented in C on the SP2 and tested on digital mammograms and other images. The results obtained are presented in this paper.