Auxiliary diagnosis of small tumor in mammography based on deep learning

As the most common cancer disease in the world, breast cancer is the main cause of death of female cancer patients. The number of women in China is far greater than that in other countries. Therefore, the absolute death toll is often high. Even a small increase in incidence rate will lead to a severe increase in deaths. Therefore, accurate preoperative diagnosis of benign and malignant breast cancer and the status prediction of various clinical indicators are urgently needed in clinical practice. Deep learning based on neural network has been widely used in diagnosis. Therefore, this paper attempts to explore the influence of deep learning on the auxiliary diagnosis technology of small tumor in mammography. In this paper, 200 cases of breast disease patients in a hospital of our city were taken as the research object, and the artificial neural network model was established. Through the experimental simulation, the results showed that the diagnostic sensitivity, specificity and overall accuracy of BP neural network for test set samples were 95.3%, 96.7 and 96.2%, respectively. The experimental results confirmed its generalization ability. In this paper, the core idea of multi feature kernel hash, combined with a variety of features and deep kernel hash network framework, constructs a new multi feature based deep learning network, which can effectively express the image features of breast tumor, and complete the task of breast micro tumor detection with good performance.