DBT micro-calcification cluster benign and malignant classification method and device based on deep learning

Abstract

The invention relates to the technical field of artificial intelligence, in particular to a DBT micro-calcification cluster benign and malignant classification method and device based on deep learning. The method comprises the steps of: obtaining sample DBT data, wherein the sample DBT data comprise sample DBT images and benign and malignant classifications of DBT micro-calcification clusters corresponding to the sample DBT images; training a classification neural network according to the sample DBT data and a stochastic gradient descent method to obtain a trained classification neural network; and fusing the classification neural network with a two-dimensional neural network and a three-dimensional neural network, splicing output features of the two-dimensional neural network and output features of the three-dimensional neural network after global average pooling to obtain feature vectors, and then obtaining benign and malignant classifications of DBT micro-calcification clusters corresponding to each sample DBT data through a full connection layer and a classification layer. The problem of inaccurate classification in the prior art is solved, and the effects of performing classification through the deep neural network and further improving the accuracy of benign and malignant classification of the micro-calcification clusters are achieved.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to a method and device for classifying benign and malignant DBT microcalcification clusters based on deep learning. Background technique [0002] The benign and malignant classification of medical images is the basic task of medical image analysis, that is, to identify the severity of the lesion area. Breast cancer is the cancer with the highest morbidity and mortality among female cancers, and early diagnosis and treatment can improve the prognosis of breast cancer. Microcalcification clusters are an important early feature of breast cancer and are of great significance for early screening of breast cancer. At present, most studies on the classification of benign and malignant microcalcification clusters focus on DM (Digital Mammography), but the images of two-dimensional imaging lack a sense of hierarchy, and it is easy to cause the problem of overlapping ...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the embodiment of the present invention provides a method and device for benign and malignant classification of DBT microcalcification clusters based on deep learning, so as to solve the problem of inaccurate classification in the prior art

Images