病理切片图像分割技术的研究

Study on Segmentation of Pathological Section Image

介绍了用图像处理及模式识别技术对显微细胞图像的自动分析和分类的方法,并针对医学图像分析中的难点,提出了基于归一化彩色空间和RGB,HSV彩色模型的分割方法:利用模式识别技术中关于特征向量空间聚类的方法实施真彩色分割.这种方式有效地利用了多维特征空间对于分割目标所提供的信息,使分割的准确性有了较大的提高,解决了图像分割过程中的单个细胞检出问题.

In accordance with the difficulty in medical image analysis, this paper puts forward two kinds of segmentation methods based on standardized colorful space and RGB and HSV colorful model： the true-color segment ation using the method about eigenveetor space cluster in pattern recognition to true-color segmentation. This method make a good use of the information supplied by multiple-dimensional feature space, thus enhances the accuracy of segmentation, and solves the problem of the selecting an individual cell out from others.