Pulmonary nodule image classification method based on information fusion safety semi-supervised clustering

Abstract

The invention discloses a pulmonary nodule image classification method based on information fusion safety semi-supervised clustering, and the method comprises the steps: firstly inputting marked and unmarked pulmonary nodule images, and carrying out the image preprocessing and feature extraction; secondly, establishing a risk degree evaluation model of the marked sample; and constructing and selecting a plurality of base clustering methods by utilizing an idea of ensemble learning, fusing division results of the base clustering methods on the marked sample through a D-S evidence theory method,and obtaining the risk degree of the marked sample according to a fusion result; then, constructing an optimization model based on graph regularization; and finally, solving the optimization model byadopting an iterative optimization strategy to obtain a clustering result. According to the method, the problem of safe use of the marked sample is solved, and the accuracy and robustness of pulmonary nodule classification are improved.

Description

technical field [0001] The invention relates to a robust semi-supervised clustering method for labeled samples, in particular to a robust semi-supervised clustering algorithm based on fusion weighting of DS evidence theory, which belongs to the field of data mining of medical images. Background technique [0002] The World Health Organization showed in the latest global cancer report in 2018: the global incidence of lung cancer (2.1 million new patients, accounting for 11.6% of the incidence of various types of cancer) and mortality (1.8 million deaths, accounting for 1.6% of all types of cancer deaths) rate of 18.4%) are the highest. Lung cancer has almost no symptoms in the early stage, and most of the patients are already in the advanced stage when they are discovered. Therefore, early detection of lung cancer is very important to improve the survival rate. Lung cancer generally exists in the form of nodules in the early stage, and timely detection of pulmonary nodules c...

AI Technical Summary

Problems solved by technology

[0004] The present invention aims at the disadvantage that the traditional classification method of pulmonary nodules based on semi-supervised clustering does not consider the risk of labeled samples, which may lead to a decline in the final classification effect, and proposes a pulmonary nodule based on information fusion security semi-supervised clustering Image Classification Methods

Images