中文摘要：

从模糊的MRI脑肿瘤的图像中准确找出肿瘤,供医学临床使用。分别采用了2种分割方法：一种是改进的阈值方法,即在进行最大方差阈值法之前首先采用手动阈值法;另一种是采用圆形结构元素作为种子的形态学分割方法。这2种方法均实现了从低对比度MRI脑图像提取目标的要求。灵活使用阈值法和形态学分割方法与医学图像中,具有一定的临床实用价值。

英文摘要：

To find tumors from the blurred MRI brain tumor images accurately for medical clinic,two segmentation methods are adopted.One is the improved threshold method： the manual threshold method is used before the implementation of maximum variance threshold method.Another is the morphological segmentation method which takes a circular structure element as the seed.Both of them can meet the requirement of extracting the tumor target from the low-contrast MRI brain images.The flexible application of threshold method and morphological segmentation method to the medical images has a certain clinical value.