Machine Learning in Medical Imaging

Resources for learning Deep and Machine Learning - esp. in the context of AI in medical applications

+++ Consider going carbon neutral with your DL efforts - don't underestimate the impact of intense network training on our climate. This is one great example https://wp.nyu.edu/ml2/carbon-neutral-lab/.

1. Useful Links to start out

• Tensorflow Homepage
• Udacity Free Course on Tensorflow
• Keras Guide A guide to the Keras API for TF which is much more high-level than raw TF or other frameworks
• ML Terminology Beginner friendly collection of terms and their explanations in DL, ML and TensorFlow
• Book on Neural Networks and Deep Learning Very good intro to NN and Deep Learning by Michael Nielsen

2. Articles discussing general concepts

• Guide to CNNs
• ReLU Explained
• Understanding Semantic Segmentation with UNET
• Techniques to prevent overfitting
• Understanding CNNs with Visualizations

3. Guides useful for medical imaging

4. Useful Implementations

• Segmentation Models with pretrained backbones in PyTorch

5. Articles on best practices, implementations and pitfalls when using AI in medicine

• Luke Oakden-Rayner's Blog A PhD student with lots of very useful blog entries discussing the impact of AI in clinical practice
• How to get clincal AI tech approved
• ImageNet Implementation in Keras
• Object Detection:
  • YOLO (v1, v2, v3) Explained
  • YOLOv3 in Keras Explained

6. Other Resources

• Kaggle vs Colab Hardware Comparision Medium post with a few useful commands and comparision of Colab vs Kaggle notebooks for various tasks
• Best GPU Cloud Providers
• Building a PC for Deep Learning
• Deep Learning Hardware Guide Tim Dettmers has a very extensive hardware guide and it seems to be updated frequently

Thanks to all the people who use their valuable time to share their knowledge and enable others to get started!