Abstract

Computed radiography (CR) images will require large storage facilities and long transmission times for picture archiving and communications system (PACS) implementation. The American College of Radiology and National Equipment Manufacturers Association (ACR/NEMA) group is planning to adopt a JPEG2000 compression algorithm in the Digital Imaging and Communications in Medicine (DICOM) standard for better utilization of medical images. The purpose of this study was to evaluate and compare the results of JPEG and JPEG2000 compressions in medical images. We applied various compression ratios of JPEG and JPEG2000 to chest, abdomen, and hand images obtained using the AGFA CR (Computed Radiography) system (AGFAMD-4.0, BELGIUM) and the FUJI CR system (FCR 9000C, JAPAN). A quantitative evaluation was carried out using the PSNR (peak signal to noise ratio), which is a commonly used measure for the evaluation of reconstructed images and measures how the reconstructed images differ from the original images. In the JPEG and JPEG2000 comparison for compression ratios up to 60:1, the PSNR value showed less than 30 dB in accordance with CR modalities and image regions in JPEG compression whereas JPEG2000 showed more than 30 dB in the same compression ratios for both AGFA and FUJI CR images. As the compression ratios were increased to 200:1, use of the JPEG2000 algorithm showed over 30 dB of the PSNR value in accordance with CR modalities and image regions. In general, when the PSNR value gets greater than or equal to 30 dB, the quality of the processed image is acceptable. Nevertheless, the PSNR value cannot absolutely stand for the quality in terms of perceptibility. The artifact of blockiness is found to be visible not only for the AGFA CR image at 40:1 but also for the FUJI CR image at 40:1 and 60:1 in JPEG compression while no significant artifact were observed in the JPEG2000 compressed image up to 200:1. In conclusion, JPEG2000 appeared to be a more effective method of compression than JPEG in medical imaging applications based on the PSNR value, ROC studies may be needed to establish whether these applications are suitable for real clinical situations requiring diagnostic accuracy.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call