Study Of A Novel Approach To Template-Based Medical Image Compression

As hospitals move toward digital age, the compression of medical image for archiving and telemedicine play a key role. On one hand, although lossy compression techniques yield high compression rates, it can cause legal problem. On the other hand, lossless compression yield low compression rates. While maintaining the quality of the image and improving the compression rates, the hybrid method which is based on clinically relevant region or region of interest ROI has become popular. Lossless compression will be applied in the clinically relevant areas and lossy compression will be applied in the other areas. The paper present a new template-based approach .which belongs to ROI-based compression method, after investigating several acceptable ROI-based compression methods in this domain. First , the templates are collected and clinically relevant areas are predefined by radiologists. Then when an image is taken of a patient, it will be aligned to one similar template in the collection so that the ROT stored in template can be assigned on the patient image. Finally the residual image which is obtained by subtracting the two image will be compressed. A key aspect to this proposal is that the process will be automated and easy to use. Physicians will not use a system that is interactive, time-consuming, or not easy to use. Of course, special manual modification will be allowed too. Another key aspect is that instead of the original image , the residual image would be compressed. It has a relatively low data after registration and subtraction, so a higher compression rate will be achieved. For better compression in implementation , here comparisons have been taken among several compression methods about non-ROI to verify the former results. The ROI are compressed using JPEG 2000 .wavelet based algorithm.