基于改进的GAC模型的医学图像序列分割算法

为了快速有效地提取出图像序列的边缘，提出了一种基于改进的测地线活动轮廓（GAC）模型的图像分割算法。在该方法中，只需在第一幅图像中感兴趣区域的内部给出大致的初始轮廓。在后续图像中，首先采用运动估计与区域统计特征结合的方法得到轮廓模型的初始轮廓，然后利用结合先验信息的测地线活动轮廓模型进行分割。此外，为了有效地减少算法运算时间，采用手工办法在第一张图像上选定模型演化的区域，该区域在后续图像上将依据分割结果自动调整大小和位置。实验结果表明：方法能够快速有效地提取目标物体的边缘。

Image segmentation algorithm for medical image sequence based on improved GAC model

In order to quickly and effectively extract the boundary of the image sequence,an image segmentation algorithm based on improved Geodesic Active Contou（rGAC）model is proposed.In the first image,the initial contour in the internal region of interest can be roughly specified by manual.There are two steps for the method in each image followed.Firstly,the initial contour is obtained by using motion estimation and statistical property.Secondly,the image is segmented by modified geodesic active contour model added by prior information.To reduce the computing time,a region for model evolution is got by manually in the first image,which can automatically adjust the size and position in other images according to the segment result.The experimental results show that the new method can quickly and effectively extract boundary of regions in medical images.