Ultrasound Image Segmentation based on the Mean-Shift and Graph Cuts Theory

This study addressed the issue of vascular ultrasound image segmentation and proposed a novel ultrasonic vascular location and detection method. We contributed in several aspects: Firstly using mean-shift segmentation algorithm to obtain the initial segmentation results of vascular images; Secondly new data item and smooth item of the graph cut energy function was constructed based on the MRF mode, then we put forward swap and α expansion ideas to optimize segmentation results, consequently accurately located the vessel wall and lumen in vascular images. Finally comparison with experts manually tagging results and Appling edge correlation coefficients and variance to verify the validity of our algorithm, experimental results show that our algorithm can efficiently combines the advantages of mean-shift and graph-cut algorithm and achieve better segmentation results.