Deep Learning

Deep Learning Resources

• Deep Learning State of the Art (2019) - MIT (Youtube Video)

• Deep Learning Fundamentals course from Lightning AI (PyTorch Lightning)

Deep Learning Applied to Medical Analysis

• A survey on deep learning in medical image analysis

• Deep Learning in Medical Imaging - Ben Glocker, Imperial College London (YouTube Video)

Deep Learning Applied to MRI

• An overview of deep learning in medical imaging focusing on MRI

• Deep convolutional neural networks for brain image analysis on magnetic resonance imaging: a review

CNNs

• How do Convolutional Neural Networks work?

• The best explanation of Convolutional Neural Networks on the Internet!

• Convolutional Neural Networks (CNN) Introduction

• A Beginner’s Guide to Understanding Convolutional Neural Networks

• CS231n Lecture Notes

U-Net

• arXiv: U-Net: Convolutional Networks for Biomedical Image Segmentation (Original Paper)

• Presentation from the Authors

• Getting acquainted with U-NET architecture + some keras shortcuts

• Satellite Image Segmentation: a Workflow with U-Net

Spinal Cord Segmentation on MRI Data

• Segmentation of the human spinal cord

• Spinal cord gray matter segmentation using deep dilated convolutions

• Automatic segmentation of the spinal cord and intramedullary multiple sclerosis lesions with convolutional neural networks

Continual Learning

• Continual Learning Papers

• Continual Learning: Towards “Broad” AI

Feature-wise Linear Modulation (FiLM)

• FiLM: Visual Reasoning with a General Conditioning Layer

  • GitHub: ethanjperez/film

  • GitHub: rosinality/film-pytorch

  • GitHub: pytorch/contrib/torchcontrib/nn/functional.py

  • GitHub: pytorch/contrib/torchcontrib/nn/modules/linear.py

• Feature-wise transformations

• Efficient Video Object Segmentation via Network Modulation

  • GitHub: linjieyangsc/video_seg

• Disentangled representation learning in cardiac image analysis

• Language Guided Fashion Image Manipulation with Feature-wise Transformations*

• Deep Multi-Class Segmentation Without Ground-Truth Labels

• GitHub: ap229997/Neural-Toolbox-PyTorch/blob/master/film_layer.py

• Hetero-Modal Variational Encoder-Decoder for Joint Modality Completion and Segmentation

GANs

• A Beginner’s Guide to Generative Adversarial Networks (GANs)

• Generative Adversarial Nets

• Towards Data Science: Understanding GANs

• Improved Techniques for Training GANs

GANs for Synthetic Medical Data

• https://ieeexplore.ieee.org/document/8653423

• arXiv: Synthetic Medical Images from Dual Generative Adversarial Networks

• arXiv: Deep Generative Adversarial Neural Networks for Realistic Prostate Lesion MRI Synthesis

• arXiv: Medical Image Synthesis for Data Augmentation and Anonymization Using Generative Adversarial Networks

• https://ieeexplore.ieee.org/abstract/document/8363576

• arXiv: GAN-based Synthetic Medical Image Augmentation for increased CNN Performance in Liver Lesion Classification

• https://link.springer.com/chapter/10.1007/978-3-319-66179-7_48

Diffusion Models

• Beginner’s Tutorial on Diffusion Models and Flow Matching

Few-shot Learning and Meta-Learning

• https://medium.com/quick-code/understanding-few-shot-learning-in-machine-learning-bede251a0f67

• Few-shot 3D multi-modal Medical Image Segmentation using Generative Adversarial Learning

• Few-Shot and Zero-Shot Multi-Label Learning for Structured Label Spaces

• Meta-Learning for Medical Image Classification

• Generalizing from Few Examples with Meta-Learning

• Few-shot Learning Using a Small-Sized Dataset of High-Resolution FUNDUS Images for Glaucoma Diagnosis

• Few-Shot Image Recognition by Predicting Parameters from Activations

Domain Adaptation:

• arXiv: Domain Adaptation for Visual Applications: A Comprehensive Survey

• arXiv: Self-ensembling for visual domain adaptation

• arXiv: Unsupervised Domain Adaptation by Backpropagation → gradient reversal layer

• Adversarial Discriminative Domain Adaptation

• JMLR: Domain-Adversarial Training of Neural Networks

• Learning Transferrable Representations for Unsupervised Domain Adaptation

• arXiv: Revisiting Batch Normalization for Practical Domain Adaptation

• Unsupervised Pixel–Level Domain Adaptation with Generative Adversarial Networks

• Adversarial Teacher-Student Learning for Unsupervised Domain Adaptation

• arXiv: Unsupervised Domain Adaptation Using Approximate Label Matching

Domain Adaptation for Medical Imaging:

• arXiv: nnU-Net: Self-adapting Framework for U-Net-Based Medical Image Segmentation

• Unsupervised domain adaptation for medical imaging segmentation with self-ensembling

• Domain Adaptation for Biomedical Image Segmentation Using Adversarial Training

• Unsupervised Domain Adaptation in Brain Lesion Segmentation with Adversarial Networks (Reference given in the Mila talk about domain adaptation in MRI)

• arXiv: Deep semi-supervised segmentation with weight-averaged consistency targets
  (auth: Julien Cohen-Adad)

• Promises and limitations of deep learning for medical image segmentation
  (auth: Julien Cohen-Adad)

• Google Books: Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support
  (auth: Julien Cohen-Adad)

• GitHub: neuropoly/domainadaptation

Neural ODEs:

• Neural Ordinary Differential Equations

• Hidden physics models: Machine learning of nonlinear partial differential equations

• Forward-Backward Stochastic Neural Networks: Deep Learning of High-dimensional Partial Differential Equations

• JMLR: Deep Hidden Physics Models: Deep Learning of Nonlinear Partial Differential Equations

Anatomical Priors:

• https://ieeexplore.ieee.org/abstract/document/8363652

• arXiv: Probabilistic Atlases to Enforce Topological Constraints

Physics Informed Deep Learning:

• arXiv: Hidden Physics Models: Machine Learning of Nonlinear Partial Differential Equations

• Yaman 2019

• PubMed: PSACNN: Pulse sequence adaptive fast whole brain segmentation

• Paperpile: Physics-Informed Brain MRI Segmentation

• arXiv: Physics Informed Deep Learning (Part I): Data-driven Solutions of Nonlinear Partial Differential Equations

• arXiv: Physics Informed Deep Learning (Part II): Data-driven Discovery of Nonlinear Partial Differential Equations

• arXiv: Hidden Fluid Mechanics: A Navier-Stokes Informed Deep Learning Framework for Assimilating Flow Visualization Data

• Label-Free Supervision of Neural Networks with Physics and Domain Knowledge

• AUTOmated pulse SEQuence generation (AUTOSEQ) and neural network decoding for fast quantitative MR parameter measurement using continuous and simultaneous RF transmit and receive

• arXiv: MR fingerprinting Deep RecOnstruction NEtwork (DRONE)

• Complex-valued Deep Learning with Applications to Magnetic Resonance Image Synthesis

• Technical Note: Simultaneous segmentation and relaxometry for MRI through multitask learning

• Knee menisci segmentation and relaxometry of 3D ultrashort echo time cones MR imaging using attention U-Net with transfer learning

• CIS522 UPenn Course: Physics Informed Deep learning

MixUp

• https://arxiv.org/pdf/1710.09412.pdf

• https://www.inference.vc/mixup-data-dependent-data-augmentation/

• https://forums.fast.ai/t/mixup-data-augmentation/22764

• http://mlexplained.com/2019/06/02/papers-dissected-mixmatch-a-holistic-approach-to-semi-supervised-learning-and-unsupervised-data-augmentation-explained/

Uncertainty

• PubMed: Improved inference in Bayesian segmentation using Monte Carlo sampling: application to hippocampal subfield volumetry

• Improving Uncertainty Estimation in Convolutional Neural Networks Using Inter-rater Agreement

• Exploring the Relationship Between Segmentation Uncertainty, Segmentation Performance and Inter-observer Variability with Probabilistic Networks

• Dropout as a Bayesian Approximation - For Epistemic Uncertainty

• What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision? - Unification of Aleatoric and Epistemic Uncertainty

• What My Deep Model Doesn’t Know…

Inter-rater

• arXiv: The Worrisome Impact of an Inter-rater Bias on Neural Network Training

• arXiv: On the Effect of Inter-observer Variability for a Reliable Estimation of Uncertainty of Medical Image Segmentation

• Evaluating White Matter Lesion Segmentations with Refined Sørensen-Dice Analysis

• Improving Uncertainty Estimation in Convolutional Neural Networks Using Inter-rater Agreement

Out of Distribution (OOD)

• Which MOoD Methods work? A Benchmark of Medical Out of Distribution Detection

• Improving Calibration and Out-of-Distribution Detection in Medical Image Segmentation with Convolutional Neural Networks

• Detection and Retrieval of Out-of-Distribution Objects in Semantic Segmentation

Longitudinal MS Lesion Segmentation

This is for the MS Challenge 2021.

• Fully automated longitudinal segmentation of new or enlarged multiple sclerosis lesions using 3D convolutional neural networks

• Longitudinal multiple sclerosis lesion segmentation: Resource and challenge