Lung CT image aided detection COVID-19 based on Alexnet network
Tao Wang, Yongguo Zhao, Lin Zhu, Guangliang Liu, Zhengguang Ma, Jianghua Zheng
Abstract
In this article, we use the Alexnet network in deep learning to determine whether the lung CT images are infected with covid-19.First of all, in the data preprocessing stage, the original CT image data is scaled and normalized to reduce noise interference. The batch operation of training set and test set can increase the training speed;Secondly, build an eight-layer Alexnet network model, set reasonable hyperparameters for each layer of the network, define the loss function and optimizer, and use the processed data to train the weight parameters of each layer in the network model.Finally, three indicators of accuracy, accuracy and recall are used to quantify the effect of model classification, and the impact of different training times on these three indicators is compared to select the best classification model.At the same time, use pyqt5 to write the corresponding auxiliary detection interface to facilitate the display of test results and the selection of classification models.The construction of the network model, the definition of the loss function and the definition of the optimizer are all based on the Pytorch deep learning framework.