Figure 2 | Scientific Reports

Figure 2

From: An artificial neural network-based radiomics model for predicting the radiotherapy response of advanced esophageal squamous cell carcinoma patients: a multicenter study

Figure 2

(a) CNN model consisting of two convolution layers with a kernel size of 3*3. Each convolutional layer follows the ReLU and max-pooling functions. The dropout function can significantly prevent the overfitting problems of the model. (b) The pretrained model was mainly built with pretrained ResNet50 from ImageNet, and the channel attention layer used both max pooling and average pooling to calculate the channel attention feature map.

Back to article page