Automatic Detection of the Intimal and the Adventitial Layers of the Common Carotid Artery Wall in Ultrasound B-Mode Images Using Snakes

This study presents a novel automatic system for detecting the intimal and adventitial layers of the far common carotid artery (CCA) by applying the snake techniques. Cohen’s snake is also modified and some criteria are added for our applications. In addition, the oscillating problem of using snakes is solved by properly choosing the time step from analysis of the frequency response of the filters. The snake scheme combined with a time diminishing gravity window and an external force makes it possible to detect the intima-media complex of the far CCA. Besides, a cost function assists the snake in selecting the optimal shape of adventitia layer. This mechanism is also helpful when there are plaques on the CCA wall. Moreover, this study compares the proposed snake and ziplock snake with respect to the manual extraction contour. According to that comparison, the system can automatically detect the intimal and adventitial layers. In addition to that it does not need any manual correction and it is noise-resistant, the proposed system allows a user to quantitatively measure an important predictor of heart infarction and strokes.