A Retinal Image Compression and Decompression Algorithm Relies on Regions of Interest

In any medical organization, there are large number of medical images are available. To keep the improvement records of the patient, it is necessary to take the medical images in regular basis. It is very much important to store those images for future references as well as a doctor's notes. These large numbers of images are needed to be stored in storage place and also to be transmitted from one storage place to another medical sector. Due to the controlled bandwidth of transmission channel and storage capacity, medical images should be compressed before storage and transmission. The privacy and security of medical images are essential. Compression should be done in such a way so that no one can access the patient data without decryption key. In this work, we have proposed a novel compression algorithm which compress the clinically relevant region of medical image as well as decryption algorithm to reconstruct original image with 100% fidelity, so that minimum space will be taken to store those images sequentially. The suggested technique is quantitatively tested against 25 eye OCT images to assess the vision loss caused by sub-retinal fluid buildup and 89 fundus images of DIARETDB1 database to find vascular structure. When compared to existing approaches in the literature, the created algorithm produces equivalent or even superior compression ratio.