Fig. 2From: Noninvasive KRAS mutation estimation in colorectal cancer using a deep learning method based on CT imagingThe structure of the employed residual neural network. There are six identity blocks, a pooling layer, a fully-connected layer and a softmax. Each identity block has three convolutional layers. The kernel size of all the convolution layers is 5 ×5. ReLu are adopted after every convolutional layerBack to article page