基于三维区域增长的脑部 MR图像分割方法研究与实现

摘要

三维脑部MR图像分割是医学图像处理中的一项重要技术，在医学临床诊疗方面有较强的学术价值和很好的应用前景，良好的分割结果可为医生的诊断和治疗提供可靠的依据。针对三维脑部MR图像的特点，首先对其进行预处理，用高斯滤波去噪得到较平滑的图像；其次，利用阈值分割和三维区域增长分割分别对三维脑部MR图像进行粗分割和精确分割提取脑组织；最后，通过形态学运算进行处理，抑制干扰性分割，同时填充目标区域内部孔洞，显著地提高了分割效果。实验结果表明，该方法能有效地对三维脑部MR图像进行分割处理，较好地提取脑组织。%3D brain MR image segmentation is an important part in medical image processing .It has high academic value and very good application prospective in clinical medicines , and the good segmentation result can provide reliable basis for doctor ’ s diagnosis and treatment . According to the characteristic of 3D brain MR image, firstly the image is pre-processed, Gaussian filter is employed to obtain a smoother image;Secondly, the threshold segmentation and 3D region-growing segmentation are used respectively for coarse segmentation and fine segmentation on 3D brain MR image to extract the brain tissue;Finally, through the processing with morphological operation , the interfering segmentations is restrained , and at the same time the holes within the target regions are filled , which significantly improves the segmentation effects.Experimental results show that this method can segment and process the 3D brain MR images effectively and extract the brain tissue preferably well .