宁夏大学学报（自然科学版）2017，Vol.38Issue(1)：27-32,6.
基于形态学滤波的标记分水岭脑肿瘤图像分割
Brain Tumor Image Segmentation of Marked-watershed Based on Morphological Filter
摘要
Abstract
The accurate segmentation of tumor in brain MRI(magnetic resonance imaging) images is usually difficult due to noise,magnetic field and volume effect,to get precise segmentation with less contour position bias,an image segmentation of marked-watershed based on morphological filter is proposed.Firstly,the morphological gradient is used to preprocess brain tumor image.Then the gradient image is filtered by opening-and-closing-by-reconstruction filter,which not only preserves contour information of brain tumor gradient image,but also removes the noise and the local minimum area.And then the extended extreme value transformation and the forced minimum technique are used to get the internal and external tags.These tags are used to correct the gradient magnitude image.Finally,the gradient image with tags is segmented by watershed algorithm.Three clinical MRI brain tumor images,which are similar to rotundity,triangle and so on and have different brightness and shape features,are simulated on segmentation test with Matlab.The experimental results show that the proposed algorithm has good segmentation accuracy and speed.关键词
图像分割/脑肿瘤图像/开闲重建/分水岭算法/标记提取/形态学Key words
image segmentation/brain tumor image/opening and closing by reconstruction/watershed algorithm/marker-extraction/morphology分类
信息技术与安全科学引用本文复制引用
王伏增,汪西原,宋佳乾..基于形态学滤波的标记分水岭脑肿瘤图像分割[J].宁夏大学学报（自然科学版）,2017,38(1):27-32,6.基金项目
国家自然科学基金资助项目(61162020) （61162020）
