论文部分内容阅读
目的设计基于几何参数的乳腺导管原位癌图像定量分析识别程序,检测识别乳腺导管原位癌。方法 200例明确诊断的乳腺导管原位癌经HE染色的组织切片,100例用于设计软件程序,100例用于检验软件。以乳腺导管原位癌筛孔型中的筛孔作为研究对象,在筛孔中提取出定量参数指标,主要包括:细胞周长、半径、面积、距离、距离比、平均灰度、并以这些几何参数为基础设计程序。结果该程序实现了感兴趣区域选定,交互式图像分割,同时计算周长、半径、面积、距离、距离比、平均灰度参数,软件可以输出并保存测量结果,对100例明确诊断病例分析,准确率为98%。结论以细胞的周长、半径、面积、距离、距离比、平均灰度为几何参数设计的图像分析程序可以检测乳腺导管原位癌(筛孔型)。
OBJECTIVE: To design a procedure for the quantitative analysis of breast duct carcinoma in situ based on geometric parameters and to detect and identify ductal carcinoma in situ. Methods 200 cases of well-differentiated ductal carcinoma in situ were stained with HE. 100 cases were used to design the software program and 100 cases were used to test the software. Taking the mesh of breast duct carcinoma in situ as the research object, quantitative parameters were extracted from the mesh, including cell circumference, radius, area, distance, distance ratio and average grayscale Geometric parameters based on the design process. Results The program achieved the selected region of interest, interactive image segmentation, and calculate the perimeter, radius, area, distance, distance ratio, average grayscale parameters, the software can output and save the measurement results, the analysis of 100 cases of clear diagnosis , The accuracy rate is 98%. Conclusion The image analysis program based on geometric parameters of cell circumference, radius, area, distance, distance ratio and average grayscale can detect ductal carcinoma in situ (mesh type).