Improved T-Snake Model Based Edge Detection of the Coronary Arterial Walls in Intravascular Ultrasound Images

Accurately detecting edges of the coronary arterial walls in intravascular ultrasound (IVUS) images plays an important role in quantitative assessments of the coronary artery diseases, though it may be a challenge to medical image analysis. This paper presented a method to extract the edges, main characteristics of which were summarized as: 1) The method was based on an improved topologically adaptable snake model (T- Snake model). This proposed T-Snake model could successfully address the significant problem of conventional T-Snakes when dealing with the model self-intersection. 2) The method worked in conjunction with an adaptive homomorphic spatial/temporal filtering technique. This proposed filtering technique was capable of effectively reducing the strong blood speckle noise in IVUS images. The experimental results indicated the proposed method was accurate and robust in detecting edges of the coronary arterial walls in IVUS images, as well as reproducible for sequential IVUS frames.