Analyses of Higher Order Metrics for SPIHT Based Image Compression

Abstract

The increasing use of digitized images has led to the need to compress such imagery to allow economical storage and fast data transfer. Despite all the advantages of JPEG compression schemes based on DCT namely simplicity, satisfactory performance, and availability of special purpose hardware for implementation; these are not without their shortcomings. The input image needs to be blocked which results in noticeable and annoying blocking artifacts particularly at low bit rates. So over the past several years, the wavelet transform has gained widespread acceptance in signal processing in general and in image compression research in particular. SPIHT codes the individual bits of the image wavelet transform coefficients following a bit-plane sequence. The evaluation of an image compression system is a difficult problem. While mean square error and peak signal to noise ratios are easily and commonly calculated, they are widely recognized to not to be completely satisfactory. The statistical parameters include high order image statistics like skewness and kurtosis which describe the shape and symmetry of the image. The aim of this paper is to provide a uniform gauge of the performance of data compression processes. In addition, it specifies how the performance of different data compression methods should be ranked so that the best compressor for a specific application can be identified. In the present work, three compression algorithms viz. the JPEG coding, the wavelet transform coding and the SPIHT coding have been discussed and compared.

Compression techniques on MR image — A comparative study

“…Jindal et al compared the SPIHT coding method with the JPEG and wavelet transform coding methods using higher order metrics via skewness and kurtosis. They found that the SPIHT algorithm performed better at lower bit rates and skewness, where as JPEG was found to have better kurtosis [10].…”