Evaluation of Mammographic Phantom Images Using JPEG2000 Image Compression

ABSTRACT

PURPOSE: To determine the usefulness of compression standard JPEG2000 for compression of mammographic images. MATERIALS AND METHODS: Image of a mammographic phantom was compressed using JPEG2000 at ratios of 101, 201, 301, 401, 501 and 601. The sizes of the images were compared, and scores were recorded by counting the numbers of fibers, groups of specks and masses seen in each phantom image. More than four fibers, three groups of specks and three masses and a total score of 10 were considered acceptable. RESULTS: The size of a DICOM image was 17,042 KB, a TIFF image was 8,324 KB, the original JPEG image was 1,506 KB and the most compressed image (501) above an acceptable total score of 10 was 43 KB. In each category, the compression image of fiber was acceptable up to compression ratio of 501 (score of 5), groups of specks was acceptable up to 601 (score of 3) and mass was acceptable up to 501 (score of 3.5). The total score, which was acquired by adding up the individual scores of all three categories, for a compression ratio of 501 was 12 and was acceptable, but the total score for 601 was 8 and was not acceptable. CONCLUSION: The compression standard JPEG2000 is an efficient means for compressing mammographic images at high ratios without compromising diagnostic value.