Automated system and method for testing resolution of ultrasound scanners

Abstract

In accordance with the present invention there is presented an automated system for testing the ability of clinical ultrasound scanners to detect focal lesions in human tissue. An ultrasound scanner phantom containing background material mimicking the ultrasonic characteristics of human tissue and coplanar spherical target lesions ultrasonically contrasting with the background material is scanned using the ultrasound scanner to be tested. Digitized images are made of the ultrasound scan of slices in the ultrasound phantom containing background material only and of slices on which the focal lesions are centered. A lesion signal to noise ratio (SNR).sub.L is then calculated at each defined pixel coordinate in the target lesion slice. This calculation employs (1) a pixel value average calculated over a sample area of the target image slice centered at a pixel location and which is of a size approximately that of the cross-sectional area of the target lesions, (2) an average pixel value calculated over an averaging area centered at the same pixel coordinate but containing mostly background image data, and (3) a standard deviation of averaged pixel values calculated in the background material image plane. The proximal and distal depth range limits of detectability of an ultrasound scanner, for a given lesion diameter and contrast, may then be determined based on the number of pixel locations in a depth range of the scanned plane having an absolute value SNR greater than a threshold value.