Clinical Evaluation of the JPEG2000 Compression Rate of CT and MR Images for Long Term Archiving in PACS.
10.3348/jkrs.2006.54.3.227
- Author:
Soon Joo CHA
1
;
Sung Hwan KIM
;
Yong Hoon KIM
;
Yoon Joon HWANG
;
Jung Wook SEO
;
Su Young KIM
;
Mi Young KIM
;
Hae Ryeon KIM
;
Yoon Hee HAN
;
Gham HUR
;
Joo Hwan PARK
;
Byung Hoon LEE
;
Seung Tae LEE
;
Bae Geun OH
Author Information
1. Inje University Image Research Center, Korea. chasoonj@ilsanpaik.ac.kr
- Publication Type:Original Article
- Keywords:
PACS;
PACS Digital imaging and communications in medicine (DICOM);
Data compression
- MeSH:
Data Compression;
Noise
- From:Journal of the Korean Radiological Society
2006;54(3):227-233
- CountryRepublic of Korea
- Language:Korean
-
Abstract:
PURPOSE: We wanted to evaluate an acceptable compression rate of JPEG2000 for long term archiving of CT and MR images in PACS. MATERIALS AND METHODS: Nine CT images and 9 MR images that had small or minimal lesions were randomly selected from the PACS at our institute. All the images are compressed with rates of 5:1, 10:1, 20:1, 40:1 and 80:1 by the JPEG2000 compression protocol. Pairs of original and compressed images were compared by 9 radiologists who were working independently. We designed a JPEG2000 viewing program for comparing two images on one monitor system for performing easy and quick evaluation. All the observers performed the comparison study twice on 5 mega pixel grey scale LCD monitors and 2 mega pixel color LCD monitors, rspectively. The PSNR (Peak Signal to Noise Ratio) values were calculated for making quantitative comparisons. RESULTS: On MR and CT, all the images with 5:1 compression images showed no difference from the original images by all 9 observers and only one observer could detect a image difference on one CT image for 10:1 compression on only the 5 mega pixel monitor. For the 20:1 compression rate, clinically significant image deterioration was found in 50% of the images on the 5M pixel monitor study, and in 30% of the images on the 2M pixel monitor. PSNR values larger than 44 dB were calculated for all the compressed images. CONCLUSION: The clinically acceptable image compression rate for long term archiving by the JPEG2000 compression protocol is 10:1 for MR and CT, and if this is applied to PACS, it would reduce the cost and responsibility of the system.