Selection of Optimum Medical Image Compression Method using Cuboid Representation

Abstract

Medical Images can be compressed using several lossy and lossless compression algorithms. In this research, brain CT images were used to study the effects of compression using EZW, SPIHT, STW, WDR and ASWDR methods and filters Haar, Db, Symlets and Biorthogonal filters. The effectiveness of the algorithms is defined by MSE, PSNR, BPP and CR. For an ideal condition MSE should be minimum, PSNR should be high, the selection of BPP should be such that, BPP should be low but at the same time the compressed image should be clearly visible. Images should be compressed at a higher compression ratio without compromise in the quality of the compressed image.
To decide on the best possible combination which helps in reduction of storage space and cost and also taking into account the quality of the compressed image, it is required to compare the huge tabulated data set. This is a tedious job. Therefore in order to help the physicians or the radiologists to choose the best combination, we have come up with a cube structure.