TODO for the image processing module

[yenchiah]

• Add data augmentation in the pipeline, ideally also showing the difference between using data augmentation and not using it
• Add an overfitting example (maybe change MNIST to another dataset that can show overfitting, such as the CIFAR-10 dataset)
• Change the ResNet graph in the notebook
• Reduce the dimension to 2D or 3D and plot the clusters using like t-SNE or PCA (so that students learn the idea of exploring clusters)
• Visualize the regions that the neural network pays attention to (so that students can learn the idea to inspect their models), using things like GradCAM or LIME
• Add an example of transfer learning (e.g., using pretrained models on other datasets), ideally also showing the difference between using pretrained models and not using it