Compression of MRI brain images based on automatic extraction of tumor region

In the compression of medical images, region interest (ROI) based techniques seem to be promising, as they can result in high ratios while maintaining quality diagnostic importance, ROI, when image is reconstructed. this article, we propose a set-up for brain magnetic resonance imaging (MRI) images on automatic extraction tumor. Our approach first separate tumor, ROI our case, from image, using support vector machine (SVM) classification and step. Then, tumor compressed Arithmetic coding, lossless technique. The non-tumorous region, non-region (NROI), lossy technique formed by combination discrete wavelet transform (DWT), set partitioning hierarchical trees (SPIHT) arithmetic coding (AC). performance parameters, like, dice coefficient, sensitivity, positive predictive value accuracy are tabulated. case compression, report, parameters like mean square error peak signal noise ratio given bits per pixel (bpp) values. We found that scheme considered setup gives promising results compared other schemes.