利用血管内超声视频图像获取冠状动脉截面积曲线

ESTIMATING THE CROSS SECTIONAL AREA OF THE CORONARY ARTERY FROM INTRAVASCULAR ULTRASOUND IMAGES

利用冠状动脉内超声的视频图像 ,依据相邻帧的连续性自动确定血管内壁的搜索范围 ,结合表征血管内壁平滑性的内部能量和表征超声图像灰阶分布的外部能量 ,采用动态规划方法通过寻找最大能量路径来检测血管内壁的最优包络 ,并由此计算血管的截面积。现提出了一种综合图像灰阶梯度和象素方差的外部能量表示 ,使得冠状动脉截面积曲线的提取方法在鲁棒性、准确性方面较仅用灰阶梯度作为外部能量有了较大的提高。

The cross sectional area of the coronary artery was computed using intra-vascular ultrasound (IVUS) video images. With this method, the search range for the inner arterial wall was firstly determined from the continuity of IVUS video frames. Then the internal and external energy was calculated to describe the smoothness of the arterial wall and the grayscale variation of ultrasound images respectively. Finally, the maximal energy path, which was the optimum contour of the arterial wall, was obtained using a dynamic programming (DP) method. In this paper, a novel form of the external energy was proposed, which integrated the gradient and variant of the grayscale image. In comparison with the traditional external energy form which only counted the image gradient, the robustness and accuracy of this method were apparently improved in the measurement of the cross sectional area of the coronary artery.