提出了基于自适应方向滤波方法来提取左心室轮廓。在噪声的干扰下，引入经平滑处理的方向滤波能够得到精确的边缘，所获得的边缘方向矢量能够在边缘跟踪时对边缘走向进行预测，同时对参数进行自适应调整。通过尽量少的人机交互，算法能够自动提取出左心室的轮廓。实验证明该算法增加了边缘提取的精度和一致性，同时显著地降低了计算复杂度。 An adaptive directional filtering method is proposed to extract the left ventricular contour. Under the interference of noise, the smoothed directional filter can be used to get the exact edge. The obtained edge direction vector can predict the edge direction while tracking the edge, and adjust the parameters adaptively. By minimizing human-computer interaction, the algorithm can automatically extract the contour of the left ventricle. Experimental results show that the proposed algorithm increases the accuracy and consistency of edge extraction and significantly reduces the computational complexity.