Classification and Recognition of Lung Image Based on CNN and Transfer Learning
This paper presents an innovative research on intelligent classification and recognition of lung images based on Deep Learning methods,and proposes a method based on Convolutional Neural Networks(CNN)and Transfer Learning,which uses classic CNN models such as VGG,InceptionV3 and ResNet as pre-trained models,and selects different Transfer Learning strategies according to the size and similarity of the datasets.This paper also uses techniques such as data augmentation,batch normalization and regularization to improve the generalization ability and stability of the model.We conduct experiments on two public lung image datasets of COVID-19 CT scans and LIDC-IDRI.The experimental results demonstrate the effectiveness and robustness of the proposed method,which can help improve the diagnostic efficiency and accuracy.
国家科技期刊平台
NSTL
万方数据
维普
