弱边缘医学超声图像目标区域的自动定位与分割

Automatic Locating and Segmenting Algorithm for Object Region of Ultrasound Images with Weak Contour

医学超声图像分割是图像处理中的一项关键技术。以胆结石超声图像为例,介绍一种新的弱边缘超声图像分割算法。首先采用基于直方图凹度分析的阈值分割方法确定Snake模型的初始蛇,再基于Snake模型结合贪婪算法对图像进行目标分割。实验结果表明该算法对弱边缘现象较为严重的医学超声图像进行目标分割时,定位准确,且分割效果良好。

Image segmentation is a key technology of image processing. Taking the gallstone ultrasound images as the example, a new segmenting algorithm for ultrasound images with weak contour was introduced. First, a primary snake nearby the object region contour is established with a threshold segmentation algorithm which is based on histogram concavity analysis, and then the edge of the gallstone is automatically detected based on the snake model and greedy algorithm. Experiment results show that the algorithm is accurate locating and effective segmenting for the ultrasound images with weak contour.