Measurement of Carotid Intima-Media Thickness in ultrasound images by means of an automatic segmentation process based on machine learning

The Intima-Media Thickness (IMT) of the Common Carotid Artery (CCA) is a reliable early indicator of atherosclerosis. Usually, it is manually measured by marking a few points on a B-mode ultrasound scan image of the CCA. By applying image segmentation techniques, the IMT can be detected along the artery length. A desirable feature of this process is the automation, avoiding the user dependence and the inter-rater variability. This work aims to find an effective segmentation method that allows the IMT measurement in an automatic way. Following this idea, this paper proposes an effective approach based on learning machines. The segmentation task is raised as a pattern recognition problem. Single Layer Feed-Forward Networks (SLFN) are designed and trained by means of the Optimally Pruned-Extreme Learning Machine (OP-ELM) algorithm to classify the pixels from a given ultrasound image, allowing the extraction of IMT boundaries. The proposed method has been tested using a set of 25 ultrasound images and several quantitative statistical evaluations have shown its accuracy and robustness.