Markov随机场与Gaussian曲线在MR图像分割中的应用

Application of Markov Random Field and Gaussian Curve in MR Image Segmentation

针对扫描的人脑组织MR图像边缘分辨率低、模糊性大的特点,本文提出了一种基于模糊Markov随机场和Gaussian曲线相结合的MR图像最佳阈值分割方法。该方法通过对图像的像素邻域属性的统计将模糊论引入其中,建立模糊Markov随机场,并利用Gaussian曲线对二维直方图最佳一维投影进行拟合,确定出图像中各脑组织的二维阈值点,在二维直方图上实现对脑组织的分割。通过实验表明,本算法能够有效提高脑组织的分辨率,对噪声的鲁棒性、结果区域的连通性相对于一维Otsu和二维Otsu算法都有了很大的提高。

For the low resolution and the large fuzziness of the organizational structure of human brain at the edge of MR image scanned, a better threshold MR image segmentation method based on fuzzy Markov random field clustering and Gaussian curves is proposed. In the algorithm, fuzzy theory is introduced into the statistics of the pixel neighborhood attributes, and fuzzy Markov random field is set up. Then the optimum one-dimensional projection histogram of two-dimensional histogram is fit with Gaussian curves and found segmentation points in each class region. Finally, image segmentation is realized in the two-dimensional histogram. Experiments show that the proposed algorithm can improve the effective resolution of the various brain tissues, and it is better than the one-dimensional Otsu method and two-dimensional Otsu method in the noise robustness and partition connectivity of the result.