27/05/2014

A MATLAB Toolbox for In Silico Evaluation of Motion Estimators for the Arterial Wall

Based on realistic mathematical motion models for the arterial wall, we have developed an in silico framework for objective evaluation of motion estimators in motion tracking of the arterial wall from B-mode ultrasound image sequences. The in silico framework consists of 13 three-cardiac-cycle simulated B-mode ultrasound image sequences of the carotid artery wall. Two methodologies have been followed to generate the image sequences: (a) by deforming real B-mode ultrasound images of healthy subjects or patients with carotid atherosclerosis and (b) by using the FIELD II software package. The generated image sequences simulate challenging, yet realistic, in vivo cases, where undesirable changes in image intensities stem from periodic mechanical phenomena in the arterial wall and/or noise interference. 

The application of a motion estimator to a simulated image sequence results in the quantitative evaluation of both accuracy and computational efficiency, thereby allowing the optimization of the algorithm for the particular clinical application and the comparative evaluation of different motion estimators. 

The in silico framework is presented in detail in:

- S. Golemati, J.S. Stoitsis, A. Gastounioti, A.C. Dimopoulos, V. Koropouli, K.S. Nikita, "Comparison of block-matching and differential methods for motion analysis of the carotid artery wall from ultrasound images," IΕΕE Trans Inf Technol Biomed, 16 (5), pp. 852-858, 2012.

- A. Gastounioti, S. Golemati, J.S. Stoitsis, K.S. Nikita, "Adaptive block matching methods for carotid artery wall motion estimation from B-mode ultrasound: in silico evaluation & in vivo application," Phys Med Biol., 58, pp. 8647-8661, 2013.

The MATLAB toolbox which packages the in silico framework and/or a (sub)set of the simulated image sequences can be made available upon request, provided that in every publication or other research work which will be produced
(A) the corresponding publications are cited and 
(B) the following statement is included in the main text: 
".. using a framework for in silico evaluation of motion estimators for the arterial wall, developed by the BioSim Laboratory of the National Technical University of Athens." 
Moreover, the toolbox and the simulated image data may not be redistributed, reposted on other web sites, or used for non research purposes without the express permission of the BioSim Director.