Comparative Analysis of SVD and Progressive SPIHT techniques for Compression of MRI and CT Images

Abstract

MRI and CT images generated by medical imaging techniques require different compression techniques with considerable image quality by reducing the storage space. This paper presents the performances two compression techniques for images i.e. Singular Value Decomposition (SVD) using singular values and wavelet transform based progressive structure of Set Partitioning In Hierarchical Trees (SPIHT). These two techniques are practiced on MRI and CT images of brain and the results of two techniques are compared. SVD with less singular value provides high Compression Ratio (CR) with less Peak Signal to Noise Ratio (PSNR), where as multi resolution based SPIHT technique provides more PSNR with better CR. These two techniques are compared with the quality metrics of PSNR, Mean Squared Error (MSE), CR and Bit Per Pixels (BPP). From the results, Wavelet based progressive SPIHT technique provides high PSNR, low MSE with better CR compared to SVD technique.

Keywords: Singular Value Decomposition, Set Partition in Hierarchical Tree, Peak Signal to Noise Ratio, Compression Ratio, Mean Squared Error