Early lung cancer diagnosis algorithm based on ResNet and Transformer
Early diagnosis of lung cancer plays a key role in improving the survival rate of pa-tients.The analysis of lung images by computer aided diagnostic system can help doctors to de-tect lesions early.In this paper,a deep learning framework combining ResNet and Transformer is proposed for the detection and classification of lung nodules in CT images,so as to realize the automatic diagnosis of early lung cancer.Firstly,data enhancement is carried out,and then ResNet's powerful feature extraction capability is utilized to obtain the deep features of the im-age.Meanwhile,Transformer is introduced to capture the long-range dependence relationship between features and enhance the models recognition ability of small changes in lung nodules.Experimental results show that the algorithm can effectively improve the accuracy and sensitivi-ty of early lung cancer detection.
early diagnosis of lung cancerComputer aided diagnosisResNetTransformerDeep learning
万方数据
维普
