Breast structure disorder identification method based on two transfer and convolutional neural network

Abstract

The invention discloses a breast structure disorder identification method based on two transfer and a convolutional neural network. The method is characterized by transferring model parameters trainedin an ImageNet data set to a target convolutional neural network to carry out initialization on model parameters thereof ( first transfer); then, carrying out first fine-tuning training (second transfer) on a target convolutional neural network model through malignant tumor images and normal breast tissue images; carrying out second fine-tuning training through class-information-known structure disorder images and normal breast tissue images in a training set; and finally, carrying out identification on class-information-unknown structure disorder and normal tissues. The method realizes identification based on the convolutional neural network and transfer learning; and compared with the existing method, the method achieves better classification effect and performance.

Description

technical field [0001] The present invention relates to automatic recognition of structural disorder lesions in mammography, in particular to a structural disorder recognition method based on deep learning (convolutional neural network and migration learning). Background technique [0002] Breast cancer is a common tumor that threatens women's physical and mental health. Early detection and treatment are the key to reducing the risk of breast cancer, and mammography is the most widely used screening method. Calcification, mass and structural disorder are the three most common X-ray abnormalities in breast cancer, and structural disorder is the most difficult lesion to detect. Among the currently used medical imaging technologies, low-dose X-ray mammography is an effective tool for early breast cancer detection, and is the most commonly used clinical detection method for female breast cancer. [0003] Structural disorder is an early and very important sign of breast cancer....

AI Technical Summary

Problems solved by technology

[0006] Since the training of convolutional neural network requires a large number of samples, it is very difficult to train a good target network model when the number of samples is insufficient.

Images