Digital compression of echocardiographic images: is it viable?

Abstract

The degree of information loss by compression of echocardiographic images utilizing the Joint Photographic Experts Group (JPEG) baseline system was evaluated using objective fidelity criteria. Standard echocardiograms were acquired digitally in real-time and stored on optical disk. 180 still frame images were processed in software using 5 preset quantization tables yielding average compression ratios of 7:1, 10:1, 17:1, 30:1 and 36:1. Objective parameters evaluated include the root mean square error (RMSE), signal-to-noise ratio (SNR), difference image mean and maximal pixel value shift. The mean pixel values and standard deviations of all images over all levels of compression were very similar. Even for ratios as high as 35:1, no significant bias was introduced in the pixel intensity values. Average RMSE ranged from 1.27 to 7.89 while average SNR ranged from 38.81 to 22.93. The data shows that JPEG compression at high ratios does not significantly alter objective image content and is a potential solution for reducing digital storage requirements.<>