A Cerebrovascular Image Segmentation Method Based on Geometrical Feature Point Clustering and Local Threshold

Background: For the cerebrovascular Digital Subtraction Angiography (DSA), how to restrain the patient motion artifact to improve the quality of subtraction image has an important effect on the clinical diagnosis. Methods: Currently, image registration is the main way to extract the blood vessels. However, there is usually massive calculation in the registration process. And it is usually only suitable for simple rigid motion artifact. Instead of registration way, a novel cerebrovascular segmentation method was proposed to extract blood vessels in this paper. In this method, the geometrical feature points of mask image and live image were firstly detected by SIFT algorithm under same restrain parameters. Secondly, the feature points were clustered and the subtraction of clustered point set was implemented. Then, the coordinates of the residual feature points were adjusted based on gray gradient. Lastly, the vessel image was segmented based on region growing and local threshold. Result: Experiments for the sequential cerebrovascular DSA images illustrate the applicability of this method. The quality of the vessel image after segmentation was satisfactory. The interdependency of geometrical feature information for both mask image and live image was adequately utilized in this new method. Conclusion: This method can provide accurate vessel image data for the clinical operation based on DSA interventional therapy.