Fig. 2.

Training and validation processes of the deep learning model based on the CT images. Accuracy (a) and cross-entropy loss (b) were plotted against the training step during the length of the training of the four-class classifier over the course of 54 steps. The red and black lines represent the training and validation processes, respectively. The cross-entropy loss was close to 0.5, while the final validation accuracy was 85.07%