Performance Analysis of Recent Algorithms for Compression of Various Medical Images

Data compression refers to a method that converts or transforms data from a characterization to a novel compressed characterization, where the distributed information remains the same by reducing the number of bits. With the increasing demand for storing and transmitting clinical scans, the shortage of sufficient storage space and communication bandwidth occurs. The terms "lossy" and "lossless compression" refer to two larger types of image compression methods. The lossy techniques loss some information at the time of compression on the other hand lossless may not loss information. The medical images have two important segments which are ROI (Region of Interest) and non-ROI. The ROI segment will be compressed using lossless technique whereas the lossy techniques will be applied on the non-ROI segment. At first one lossless and one lossy technique Lempel-Ziv-Welch (LZW) and Singular Value Decomposition (SVD) is implemented respectively. Then three combination of hybrid technique i.e. lossless and lossy image compression technique Huffman + Singular Value Decomposition (SVD), Lempel-Ziv-Welch (LZW) + Discrete Cosine Transform (DCT) and Set Partitioning in Hierarchical Trees (SPHIT) + Discrete Wavelet Transform (DWT) is implemented respectively. The image compression techniques are implemented in MATLAB and results is analysed in terms of Peak Signal-to-Noise Ratio (PSNR), Mean Squared Error(MSE), Structural Similarity Index Measure (SSIM) and Compression Ratio (CR). In this SPHIT + DWT shows better result in terms of PSNR and CR while compared to other techniques.