An Efficient Hybrid Image Compression Scheme using DWT and ANN Techniques

Abstract

High rate data compression is an essential task in the present world of digital communication. Individually the wavelet transform (WT) based compression method is able to provide a compression ratio of about 20–30, which is not adequate for many practical situations. The main motivation of this article is to propose a novel hybrid approach which would offer higher compression ratio than the WT alone keeping the quality of reproduced image identical in both cases. This is achieved with the incorporation of a second image compressor in sequence with WT so that the overall compression becomes better than the of individual method. The second compressor used in the paper is based on Artificial Neural Network (ANN). The WT is employed to achieve the first stage of compression, which is followed, by a second stage of compression using the ANN technique. In the ANN technique both multi layered ANN (MLANN) and the radial basis function (RBF) networks have been proposed. The compression performance has been assessed in terms of peak-signal-to-noise-ratio (PSNR) and energy retained in the reconstructed image. Through computer simulation it has been demonstrated that the combined approach offers high rate of compression maintaining identical reconstructed image quality compared to its WT counterpart.

Indexing terms:

• Hybrid image compression
• Digital communication
• WT
• Wavelet transform
• Artificial Neural Network