Wavelet compression of ultrasound video streams for teleradiology

Future developments in teleradiology hinge on the delivery of real or near real-time images, sometimes across less than optimal bandwidth communication channels. Ultrasound, to achieve its greatest diagnostic value, needs to transmit not just still images but video as well. A significant amount of compression, however, may be required to achieve near real-time video across limited bandwidths. This will inevitably result in degraded video quality. A variety of compression algorithms are in widespread use including H.261, H.323, JPEG (Joint Photographic Experts Group), MPEG (Motion Picture Expert Group) and most recently wavelets. We have developed a suite of tools to evaluate each of these methods, and to identify potential areas where wavelet compression may have an advantage. In this particular study, we compare motion wavelet compression to motion JPEG compression using the standard correlation coefficient and the normalized mean squared error, and found the motion wavelet technique slightly better.