Region-Based Lossless Image Compression for MRI Images

Medical images like MRI produces high-resolution scans, which require a lot of space and a lot time for transmission. Existing lossless compression method retains diagnostic information but they fail to provide high compression. Our work mainly uses Region-Based lossless compression technique where the input is MRI image which is first segmented into ROI (Region of Interest) and non-ROI segments. Wherein ROI part is kept without any loss, while the non-ROI is compressed more aggressively to reduce the size. This technique integrates selection of ROI box, segmentation, classification, DWT and SPIHT encoding for effective compression and reconstruction. In our project the sample of MRI image of 113.3KB was reduced to 34.3KB while retaining the complete diagnostic quality in the ROI. Other features include highlighting the ROI along with automatic report for the parameters of ROI. Results confirms that the technique offers high compression with preserved diagnostic structure, hence suitable for telemedicine and medical purpose.