Optimized GPU Framework for Ultrasound Strain Imaging

Abstract

A GPU framework for ultrasound strain image based on the motion tracking with phase zero method and strain estimation by least squares method is presented. The required signal processing is computationally intensive, involving crosscorrelation, box filter, and sum of squares, etc. In this paper, we describe the use of a graphics processing unit for implementing image processing algorithms of strain imaging that achieves a frame rate of 15 fps for the E-mode I/Q signal with 512 samples and 225 beam lines, which is about 20 times faster than the CPU implementation. Testing results from GPU and CPU are compared in terms of visual image quality and program runtime from different image sizes.