The goal of this paper is to investigate and evaluate image quality, based on quality metrics and visual perception in ultrasound imaging under different imaging conditions. We first generate and simulate Bmode ultrasound images of various objects, using Field-II simulation toolbox [1]. Then we implement and embed front-end functional modules, mid-end functions including beamforming, receive beamforming, envelop detection and log compression, and back-end image processing methods (filtering and image enhancement techniques). Ultrasound images are evaluated as pairs using various image quality evaluation metrics and visual perception evaluation. The experimental results of this study show that: (1) better image quality is significantly obtained using over-sampling and signal emphasizing (high-signal level); (2) normalized images with speckle filters are rated better in terms of index-quality; and (3) enhanced images show better visual perception on all simulation datasets. This paper shows the utility of our MATLAB test-bench favoring simulated image quality and further demonstrates that the evaluation design is an important pre-processing step especially for the hardware design of ultrasound system.

Date Published: 29 March 2013
PDF: 6 pages
Proc. SPIE 8675, Medical Imaging 2013: Ultrasonic Imaging, Tomography, and Therapy, 867516 (29 March 2013); doi: 10.1117/12.2007050

Published in SPIE Proceedings Vol. 8675:
Medical Imaging 2013: Ultrasonic Imaging, Tomography, and Therapy
Johan G. Bosch; Marvin M. Doyley, Editor(s)