HCC pathological image-oriented cell nucleus segmentation and classification method

Abstract

The invention relates to an HCC pathological image-oriented cell nucleus segmentation and classification method. The method comprises the steps of reading an original HCC image; performing k-means clustering on the image to obtain segmented cell nucleuses; performing refinement on the segmented cell nucleuses by using morphological operation, performing registration in three aspects, and performing calculation through four similarity parameters together with a cell nucleus shape library manually selected by a pathologist to obtain a cell nucleus shape feature matrix; according to the registered cell nucleuses, calculating cell nucleus boundary characteristics to obtain a cell nucleus boundary characteristic matrix; and fusing the cell nucleus shape feature matrix with the cell nucleus boundary characteristic matrix, and performing classification in a random forest classifier to obtain a result. The cell nucleuses are segmented by using the k-means clustering and the morphological operation; and the cell nucleuses are classified and identified through the proposed shape and boundary characteristics, so that the classification accuracy of each type of cell nucleuses is improved, theconditions of more cells are considered, and higher pertinence is achieved.

Description

Technical field [0001] The present invention relates to an image processing technology, in particular to a method for segmentation and classification of nuclei of HCC pathological images. Background technique [0002] Image segmentation and classification are the basic operations in image processing. The traditional method is to process images through a series of image segmentation functions of Matlab to achieve the purpose of segmentation. Subsequently, feature extraction is performed on the target object, and the specificity of the feature is used to classify the object. Although with the continuous advancement of various aspects of research, image segmentation and classification have made great progress, there is still a lot of room for improvement in the accuracy of pathological image segmentation in medicine, especially for the segmentation and classification of cancer cells. . Liver cancer is the world’s second most fatal cancer. In clinical diagnosis, early recognition o...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the disadvantages of the prior art, such as difficult segmentation, low classification accuracy, and lack of pertinence in the segmentation of cell nuclei in HCC pathological images, the problem to be solved by the present invention is to provide a method that can realize rapid segmentation of pathological images, improve A Segmentation and Classification Method for Nuclei in HCC Pathological Image Based on the Accuracy of Cell Nuclei Classification

Images