Differential Coding-Based Medical Image Compression

Modern trends of technology face the challenges of cost-effective massive data storage and transmission. Image compression is a master key for this issue. Basically, the process of image compression reduces redundant and irrelevant information from the original data resulting in reduced data file size. Vector quantization is a lossy image compression technique which helps to achieve higher compression with less computation complexity. The aim of the proposed work is to develop a novel medical image compression method that blends differential encoding and vector quantization (VQ). The basic idea is to transform the input image blocks into a set of difference vectors (difference between the each pixel intensity value and its respective mean). The difference vectors are normalized to preserve the sign and further quantized to generate the codebook. The algorithm is also investigated with other statistical moments like median and mode for finding the difference vectors. The experimental results with test medical images have demonstrated better performance of the proposed method when compared to similar methods.