An Enhanced Method for Automatic Detection and Segmentation of Carotid Artery in Ultrasound Images
Lačezar Ličev, Karolina Feberova, Jan Tomeček, Jakub Hendrych · 2016
This paper deals with processing of ultrasonic images of carotid arteries. These images often suffer from low quality, intensity inhomogeneity and artefacts. Due to this, edges of segmented objects may be corrupted. Conventional methods for image processing are therefore insufficient. The newly developed method suited for automatic circular object detection and segmentation is capable to keep the original arterial borders, while precisely segmenting the eventual pathology. Hough transform was used for artery detection. Active contour based on level set methods was adjusted according to our needs and applied for the object segmentation.