Digital compression of echocardiograms: impact on quantitative interpretation of color Doppler velocity.

J. D. Thomas*, S. Chandra, T. H. Karson, M. Pu, P. M. Vandervoort

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

26 Scopus citations

Abstract

To test the impact of Joint Photographic Expert Group (JPEG) compression on the quantitative data encoded in color Doppler echocardiographic images, digital images from transesophageal echocardiography and an in vitro model of proximal flow convergence were analyzed before and after JPEG compression with compression ratios (CRs) as high as 65:1. Even at the highest CRs, greater than 95% of the pixels were categorized correctly as representing structure (gray scale) and greater than 98% were categorized correctly as representing velocity (color) data. Furthermore, the velocities and flows recovered from the compressed images agreed well (r = 0.998 [velocities] and r = 0.998 [flows] for CR = 7:1, falling to r = 0.881 [velocities] and r = 0.930 [flows] at CR = 65:1; p < 0.001 for the linear trend with CR). There was similarly little shift in the location of the red-blue aliasing contour, rising from an error of 0.05 +/- 0.19 (mean +/- SD) mm at CR = 7:1 to a maximum error of 0.11 +/- 0.36 mm at CR = 44:1. Thus JPEG compression has little impact on the quantitative velocity data encoded within color Doppler echocardiograms, which should allow widespread acceptance of digital transmission and storage.

Original languageEnglish (US)
Pages (from-to)606-615
Number of pages10
JournalJournal of the American Society of Echocardiography : official publication of the American Society of Echocardiography
Volume9
Issue number5
DOIs
StatePublished - Jan 1 1996

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging
  • Cardiology and Cardiovascular Medicine

Fingerprint Dive into the research topics of 'Digital compression of echocardiograms: impact on quantitative interpretation of color Doppler velocity.'. Together they form a unique fingerprint.

Cite this