论文部分内容阅读
目的本文试图对显微镜下病理切片的数字图像进行色度学上的分析,采用数字图像处理和图像识别的手段,初步建立对病理性核分裂像进行自动识别的量化体系。方法通过对19幅不同类型的浸润性肿瘤被不同程度染色的彩色细胞涂片图像进行分析,得到病理性核分裂像与其他正常细胞的颜色差异和量化特征,并构造相应的判别函数,对图像中病理性核分裂像进行区分。结果实验结果表明,本文提出的特征及判别函数能够在原图中有效地标示病理性核分裂像,并实现了PMF的自动计数功能。结论色度学特征可作为细胞涂片图像中区分病理性核分裂像的分类依据之一。
Objective This paper attempts to analyze the digital images of pathological sections under the microscope by means of digital image processing and image recognition to establish a quantitative system for the automatic identification of pathological mitotic images. Methods The color difference and quantification of pathological mitosis and other normal cells were analyzed by analyzing the smear images of 19 different types of infiltrating tumors which were stained with different degrees. Corresponding discriminant functions were constructed. Pathological mitosis like to distinguish. Results The experimental results show that the proposed features and discriminant functions can effectively label pathological mitotic images in the original image and realize the automatic counting function of PMF. Conclusion Chromatographic features can be used as a classification basis for distinguishing pathologic mitosis in cell smears.