Diffraction pattern recognition method for drastically improving image resolution and diagnostic sensitivity.

Abstract

Compared with the two other mainstream medical imaging methods, CT and MRI, the worst weakness of ultrasound imaging is poor resolution—the ability to resolve tiny variations in tissue structure or texture. Due to sound wave diffraction and the way imaging ultrasound signal is retrieved, a point target in object domain does not generate point image in image domain, but an image spot with sophisticate spot pattern and considerable spot size as functions of point target location relative to the transducer—the most direct cause of poor resolution. The mainstream techniques aimed at reducing the image spot, either by shortening the impulse signal or by sharpening the beam focusing, are unavoidably accompanied by the deteriorated sensitivity and depth of penetration. It is a common consent in medical ultrasound community that ultrasound imaging is very close to its theoretical resolution limit. This article presents a different approach we named E-mode imaging that uses a diffraction-theory-based spot pattern recognition technique to account for the effects of image spot pattern in image processing. With the same set of ultrasound data that B imaging is based on, E-mode imaging achieved five to ten times better resolution and diagnostic power in computer simulations.