Cervical cancer is a common malignant tumor that seriously threatens women's health.So early screening is very important.In order to solve the problem of uneven distribution of high-quality medical resources in China,the study analyzes the staining normalization and segmentation techniques of cytopathological images with Al-assisted analysis.The improved Macenko staining normalization method improves the image consistency,and the improved Mask R-CNN model enhances the segmentation performance in the case of cell overlap and masking.The results of the experiment show that these improvement significantly improve the accuracy and stability of image segmentation.
维普
万方数据