dFuse: An Optimized Compression Algorithm for DICOM-format Image Archive

Medical images are useful for knowing the details of the human body for health science orremedial reasons. DICOM is structured as a multi-part document in order to facilitate extension ofthese images. Additionally, DICOM defined information objects are not only for images but alsofor patients, studies, reports, and other data groupings. More information details in DICOM,resulted in large size, and transferring or communicating these files took lots of time. To solvethis, files can be compressed and transferred. Efficient compression solutions are available andthey are becoming more critical with the recent intensive growth of data and medical imaging. Inorder to receive the original and less sized image, we need effective compression algorithm.There are different algorithms for compression such as DCT, Haar, Daubechies which has itsroots in cosine and wavelet transforms. In this paper, we propose a new compression algorithmcalled “dFuse”. It uses cosine based three dimensional transform to compress the DICOM files.We use the following parameters to check the efficiency of the proposed algorithm, they are i) filesize, ii) PSNR, iii) compression percentage and iv) compression ratio. From the experimentalresults obtained, the proposed algorithm works well for compressing medical images.