基于边界跟踪算法的脑磁共振图像分割

目的：脑磁共振图像的自动分割是近几年研究的一大热点，本文在通过分析比较当前各种图像分割算法后，介绍了一种基于边界跟踪的脑磁共振图像（MRI）分割算法，在MRI中提取出脑组织部分。方法：应用迭代法对脑磁共振图像进行二值化处理；扫描二值化图像，根据脑组织的形态，确定一点作为脑组织边界的起点；根据边界点的像素特征，对传统的边界跟踪算法进行改进，计算出MRI脑组织的边界，最后应用区域生长法在原始MRI中提取脑组织图像，实现MRI分割。结果：实验结果表明，改进后的边界跟踪算法在提取脑组织边界时，细节处理能力强，不易陷入死循环，具有较高的运算速度。提取的真实脑磁共振图像的脑组织区域，能满足临床的实际需要。结论：对图像二值化处理，对图像有微弱的损害，但是极大地提高了计算速度。与传统的边界跟踪算法相比，改进后的边界跟踪算法分割效率高，更易实现MRI的自动分割。获得的边界曲线在细节上更接近于脑组织的实际边界。

MRI Image Segmentation Based on Edge Tracing Algorithm

Objective： The automatic segmentation of magnetic resonance images is a research hotspot in recent years. Through comparing and analyzing all kinds of image segmentation algorithms, this paper introduced a magnetic resonance image segmentation algorithm based on edge tracing, then extracted the brain tissue from the magnetic resonance image. Methods： The magnetic resonance image was transformed into a binary image by using iterative method;According to the form of the brain tissue, the starting point of the brain tissue edge was defined by scanning the binary image; According to the eigenvalue of the edge ,the traditional edge tracing algorithm was improved, and the edge of the brain tissue was found by using the improved algorithm;Finally the brain tissue was extracted from the magnetic resonance image by using region growing method. Results： The experiment results showed that the improved edge tracing algorithm could find out the detail of the magnetic resonance image, and could avoid the infinite loop on the processing of image segmentation. The experiment results also showed the fast calculation speed of the algorithm. This improved algorithm could extract the brain region perfectly and satisfy the demand of clinicalpractice. Conclusions： The method of image binarization can improve the calculation speed, but it also can damage the magnetic＇resonance image slightly. Compared with the traditional edge tracing algorithm, the improved algorithm is efficient and is easy to automatic segmentation. The image segmentation result is to be accord with the real brain tissue.