Effects of MPEG compression on the quality and diagnostic accuracy of digital echocardiography studies

Citation
K. Spencer et al., Effects of MPEG compression on the quality and diagnostic accuracy of digital echocardiography studies, J AM S ECHO, 13(1), 2000, pp. 51-57
Citations number
11
Categorie Soggetti
Cardiovascular & Respiratory Systems
Journal title
JOURNAL OF THE AMERICAN SOCIETY OF ECHOCARDIOGRAPHY
ISSN journal
08947317 → ACNP
Volume
13
Issue
1
Year of publication
2000
Pages
51 - 57
Database
ISI
SICI code
0894-7317(200001)13:1<51:EOMCOT>2.0.ZU;2-S
Abstract
The advantages of digital echocardiography studies include ease of retrieva l, review, comparison, duplication, and quantitation as well as the potenti al for moving studies over networks. However, the large amounts of data ass ociated with digital echocardiography studies have posed new problems. Redu ction of the amount of data can be accomplished with image compression, in particular MPEG-1 (Moving Pictures Expert Group), which is designed for dyn amic image sequences. However the effects of different levels of compressio n on the quality and diagnostic content of echocardiographic images need to be established. Digital sequences of single cardiac cycles were acquired in 40 consecutive patients, MPEG-1-compressed at different effective ratios (60:1, 80:1, 120: 1, 200:1, 300:1, 370:1, 500:1), reviewed, and scored for endocardial visual ization. The overall visualization scores and percentages of nonvisualized segments in the compressed images were not different from the uncompressed images up to a compression ratio of 200:1. Differential effects of compress ion were noted on a segmental basis and also varied with baseline image qua lity. The ability to correctly identify regional wall motion abnormalities did not decrease until compression ratios of 1:200 or higher were used. Digital echocardiography loops, MPEG-1-compressed at an effective ratio of 200:1, demonstrate no degradation in endocardial visualization quality or d iagnostic content. Compression at this ratio has the potential to reduce th e storage size of digital echocardiography studies to less than 1% of their current size.