CAROLAB A platform to analyze carotid ultrasound data

The CAROLAB software platform is introduced to analyze B-mode ultrasound image sequences of the carotid artery. The main objective of CAROLAB is to enable automatic, accurate, and reproducible measurements of clinical data in the context of cardiovascular risk evaluation. In addition to the most common image-based vascular biomarkers such as intima-media thickness and diameter variation, a special attention is paid to the analysis of rather unexplored phenomena, such as the shearing motion of the vessel layers in the direction of the blood flow and the compression-decompression of the IMT over the cardiac cycle. The three main methodologies currently present in CAROLAB are: 1) a contour segmentation approach based on dynamic programming to extract the anatomical interfaces of the wall; 2) a Kalman-based single-point motion-tracking approach to quantify the wall longitudinal excursion; and 3) a dense-field motion-estimation approach relying on dynamic programming to assess the motion (in-)homogeneity. All the parameters derived from these previously validated methodologies have been demonstrated to be associated with cardiovascular risk factors.