Edge extraction of the coronary arterial walls in IVUS images would be a challenge to medical image analysis,which represents a unique tool to quantitative assessments of the coronary artery diseases by extracting edges accurately.The poor quality of IVUS images makes unsupervised segmentation based on traditional segmentation algorithms fail to achieve the expected results.To extract edge automatically in intravascular ultrasound images,a novel method is proposed.Firstly,in the process of initial edge extraction,the noise and the artifacts could be reduced by using the temporal features and prior knowledge; Secondly,the new regulatory factor and adaptive external force are introduced in the GVF-Snake algorithm,it can y enlarge the edge capture range,improve active contour to noises robustness and make the extraction effect of the weak edge image more accurate.Thirdly,using cubic B-spline can improve the edge smoothness,and speed up the convergence.