some resources on my path in deep learning and medical image analysis
| Libraries | Usage |
|---|---|
| Albumentations | Fast image augmentation library and an easy-to-use wrapper around other libraries. |
| ANTsPy | Advanced Normalization Tools in Python |
| DIPY | A python library for the analysis of MR diffusion imaging |
| Gym | A toolkit for developing and comparing reinforcement learning algorithms |
| CuPy | NumPy & SciPy for GPU |
| Kornia | Open source differentiable computer vision library for PyTorch |
| Matplotlib | A comprehensive library for creating static, animated, and interactive visualizations in Python. |
| MedPy | Medical image processing in Python |
| MONAI | AI Toolkit for Healthcare Imaging |
| NumPy | Fundamental package for scientific computing with Python. |
| Optuna | A hyperparameter optimization framework |
| Psutils | Cross-platform lib for process and system monitoring in Python |
| PyTorch | Tensors and dynamic neural networks in Python with strong GPU acceleration |
| PyTorch3d | PyTorch3D is FAIR's library of reusable components for deep learning with 3D data |
| PyTorch Geometric | Geometric deep learning extension Library for PyTorch |
| PyTorch Optimizer | Collection of optimizers for PyTorch |
| PyVarInf | Facilities to easily train your PyTorch neural network models using variational inference. |
| PyVista | 3D plotting and mesh analysis through a streamlined interface for the VTK |
| SciencePlots | Matplotlib styles for scientific plotting |
| Seaborn | A Python data visualization library based on matplotlib |
| SimpleITK | A layer built on top of the ITK |
| SciPy | An open-source software for mathematics, science, and engineering |
| TensorRT | TensorRT is a C++ library for high performance inference on NVIDIA GPUs and deep learning accelerators. |
| Torchinfo | View model summaries in PyTorch |
| TorchIO | Medical image preprocessing and augmentation toolkit for deep learning. |
| Vedo | A python module for scientific analysis of 3D objects based on VTK and Numpy |
| Conference | Abbr |
|---|---|
| AAAI Conference on Artificial Intelligence | AAAI |
| ACM International Conference on Multimedia | ACM MM |
| Asian Conference on Computer Vision | ACCV |
| British Machine Vision Conference | BMVC |
| Conference on Computer Vision and Pattern Recognition | CVPR |
| Conference on Neural Information Processing Systems | NeurIPS |
| European Conference on Computer Vision | ECCV |
| International Conference on Computer Vision | ICCV |
| International Conference on Learning Representations | ICLR |
| International Conference on Machine Learning | ICML |
| International Conference on Medical Image Computing and Computer Assisted Intervention | MICCAI |
| International Joint Conference on Artificial Intelligence | IJCAI |
| International Symposium on Biomedical Imaging | ISBI |
| Medical Imaging with Deep Learning | MIDL |
Machine Learning
Overviews
- A Gentle Introduction to Graph Neural Networks
- A Survey of Convolutional Neural Networks: Analysis, Applications, and Prospects
- An Overview of Deep Learning for Curious People
- Interpretable Machine Learning
- Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges
- Loss Functions for Image Segmentation
- Spinning Up in Deep RL
Tricks
- A Recipe for Training Neural Networks
- Bag of Tricks for Image Classification with Convolutional Neural Networks
- Bag of Tricks for Training Deeper Graph Neural Networks: A Comprehensive Benchmark Study
- Deep Learning for Medical Image Segmentation: Tricks, Challenges and Future Directions
- My Neural Network isn't working! What should I do?
Basic Math
Models and Algorithms
- Anomaly Detection Using Principal Component Analysis (PCA) Reconstruction Error
- Bootstrap Confidence Intervals
- Diffusion models from scratch in PyTorch
- Flow-based Deep Generative Models
- From AE to Beta VAE
- Intuition for the Maximum Mean Discrepancy two-sample test
- Mixture Models and the EM Algorithm
- Normalizing Flows Are Not Magic
- Policy Gradient Algorithms
- Procrustes analysis with Python/NumPy
- Transforms and Resampling in SimpleITK
- Types of Distance Measures
Writing
- Aims and Objectives – A Guide for Academic Writing
- Free PDF White Space Cropping Website
- Latex Symbol and Mark (Chinese)
- Latex Table Generator
- Split PDF
- Thesaurus and Word Tools for Your Creative Needs
Research
- Best Practices and Scoring System on Reviewing A.I. based Medical Imaging Papers: Part 1 Classification
- Research Methods Knowledge Base
- Tips for Academic Research
- A Multi-site Dataset for Prostate MRI Segmentation
- Cohort Size: 116
- Modality: T2-weighted MRI
- Annotation: Segmentation of Prostate
- ACDC: Automated Cardiac Diagnosis Challenge
- Cohort Size: 150 (100 annotated)
- Modality: cine-MRI; 3D+t; short-axis
- Annotation: segmentation of LV, Myo, and RV in ED and ES; diagnosis results
- Aorta Anatomy Synthetic Dataset
- Cohort Size: 3000 (Generated from model, not from real patient)
- Modality: Mesh
- Annotation: none
- AVT: Aortic Vessel Tree CTA Datasets and Segmentations
- Cohort Size: 56
- Modality: CTA
- Annotation: segmentation of the aortas and aortic vessel trees
- CAMUS: Cardiac Acquisition for Multi-structure Ultrasound Segmentation
- Cohort Size: 500
- Modality: Ultrasound
- Annotation: Endocardium and epicardium of the LV and RV
- CMRxMotion: Extreme Cardiac MRI Analysis Challenge under Respiratory Motion
- Cohort Size: 100 (80 annotated)
- Modality: cine-MRI; ED and ES; short-axis
- Annotation: segmentation of LV, Myo and RV; motion artefact level
- EchoNet-Dynamic: A Large New Cardiac Motion Video Data Resource for Medical Machine Learning
- Cohort Size: 10030
- Modality: apical-4-chamber echocardiography videos
- Annotation: clinical measurements
- IntrA: 3D Intracranial Aneurysm Dataset for Deep Learning
- Cohort Size: 103 real patients (116 aneurysm); 1909 virtual patients (1694 healthy vessel; 215 aneurysm)
- Modality: Mesh
- Annotation: segmentation of aneurysm and vessel
- M&Ms: Multi-Centre, Multi-Vendor & Multi-Disease Cardiac Image Segmentation Challenge
- Cohort Size: 345 (320 annotated)
- Modality: cine-MRI; 3D+t; short-axis
- Annotation: segmentation of LV, Myo, and RV in ED and ES
- M&Ms-2: Multi-Disease, Multi-View & Multi-Center Right Ventricular Segmentation in Cardiac MRI
- Cohort Size: 360 (320 annotated)
- Modality: cine-MRI; 3D+t; short-axis
- Annotation: segmentation of LV, Myo, and RV in ED and ES
- MM-WHS: Multi-Modality Whole Heart Segmentation
- Cohort Size: 60 (20 annotated)
- Modality: MRI; CT
- Annotation: segmentation of four chambers, myocardium, aorta and artery.
- Virtual Cohort of Adult Healthy Four-chamber Heart Meshes from CT Images
- Cohort Size: 20
- Modality: Mesh
- Annotation: segmentation of four chambers, valve planes, aortic root, etc.
Generative Models
AutoEncoder
- Auto-Encoding Variational Bayes. ICLR 2014
- A Recurrent Latent Variable Model for Sequential Data. NIPS 2015
- beta-VAE: Learning Basic Visual Concepts with a Constrained Variational Framework. ICLR 2017
- Preventing Posterior Collapse with delta-VAEs. ICLR 2019
- Sparse Multi-Channel Variational Autoencoder for the Joint Analysis of Heterogeneous Data. ICML 2019
- Variational Laplace Autoencoders. ICML 2019
- NVAE: A Deep Hierarchical Variational Autoencoder. NIPS 2020
- Generation of Realistic Synthetic Data using Multimodal Neural Ordinary Differential Equations. npj Digital Health 2022
- Dynamical Deep Generative Latent Modeling of 3D Skeletal Motion. IJCV 2022
Flow
- Variational Inference with Normalizing Flows. ICML 2015
- NICE: Non-linear Independent Components Estimation. ICLR 2015
- Density Estimation using Real NVP. ICLR 2017
- f-VAEs: Improve VAEs with Conditional Flows. arXiv 2018
- Glow: Generative Flow with Invertible 1x1 Convolutions. NeurIPS 2018
- Learning Likelihoods with Conditional Normalizing Flows. arXiv 2019
- C-Flow: Conditional Generative Flow Models for Images and 3D Point Clouds. CVPR 2020
- Poisson Flow Generative Models. NeurIPS 2022
- ManiFlow: Implicitly Representing Manifolds with Normalizing Flows. 3DV 2022
- FLAG: Flow-based 3D Avatar Generation from Sparse Observations. CVPR 2022
GAN
- Generative Adversarial Networks. NIPS 2014
- A Style-Based Generator Architecture for Generative Adversarial Networks. CVPR 2019
- Analyzing and Improving the Image Quality of StyleGAN. CVPR 2020
- Alias-Free Generative Adversarial Networks. NeurIPS 2021
- On the "Steerability" of Generative Adversarial Networks. ICLR 2020
- LatentSwap3D: Semantic Edits on 3D Image GANs. arXiv 2022
Diffusion Model
- Denoising Diffusion Probabilistic Models. NeurIPS 2020
- Denoising Diffusion Implicit Models. ICLR 2021
- High-Resolution Image Synthesis with Latent Diffusion Models. CVPR 2022
- Diffusion Autoencoders: Toward a Meaningful and Decodable Representation. CVPR 2022
Cardiovascular
- Linking Statistical Shape Models and Simulated Function in the Healthy Adult Human Heart. PLoS Computational Biology 2021
- Clinically-Driven Virtual Patient Cohorts Generation: An Application to Aorta. Frontiers in Physiology 2021
- The Health Digital Twin to Tackle Cardiovascular Disease — A Review of An Emerging Interdisciplinary Field. npj Digital Medicine 2022
- Semi-Automated Construction of Patient-Specific Aortic Valves from Computed Tomography Images. Annals of Biomedical Engineering 2023
- Weakly Supervised Inference of Personalized Heart Meshes based on Echocardiography Videos. MedIA 2023
Geometric Learning
Generative and Editable Model
- Generating 3D Faces using Convolutional Mesh Autoencoders. ECCV 2018
- Fully Convolutional Mesh Autoencoder using Efficient Spatially Varying Kernels. NeurIPS 2020
- SceneGen: Generative Contextual Scene Augmentation using Scene Graph Priors. arXiv 2020
- Contextual Scene Augmentation and Synthesis via GSACNet. arXiv 2021
- Learning to Generate 3D Shapes from a Single Example. SIGGRAPH 2022
- Deep Deformable 3D Caricatures with Learned Shape Control. SIGGRAPH 2022
Statistical Shape Model
- Automatic Construction of Multiple-object Three-dimensional Statistical Shape Models: Application to Cardiac Modeling. IEEE-TMI 2002
- Building 3-D Statistical Shape Models by Direct Optimization. IEEE-TMI 2010
Trustworthy AI
- FUTURE-AI: Guiding Principles and Consensus Recommendations for Trustworthy Artificial Intelligence in Medical Imaging. arXiv 2021
- Guidelines and Evaluation of Clinical Explainable AI in Medical Image Analysis. MedIA 2023
- Holding AI to Account: Challenges for the Delivery of Trustworthy AI in Healthcare. arXiv 2022
Plausibility
- Learning Deformable Registration of Medical Images with Anatomical Constraints. Neural Networks 2020
- Deep Autoencoding Gaussian Mixture Model for Unsupervised Anomaly Detection. ICLR 2018
Network Interpretation
- InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets. NIPS 2016
- Understanding the Role of Individual Units in a Deep Network. PNAS 2020
- Unsupervised Discovery of Interpretable Directions in the GAN Latent Space. ICML 2020
- Inverse-Based Approach to Explaining and Visualizing Convolutional Neural Networks. IEEE-TNNLS 2022
Confidence
Robust Deep Learning
Gradient Operation
- Gradient Surgery for Multi-Task Learning. NeurIPS 2020
- Domain Generalization via Gradient Surgery. ICCV 2021
Style Generalization
- Domain Generalization with MixStyle. ICLR 2021
- A Simple Feature Augmentation for Domain Generalization. ICCV 2021
- Adversarially Adaptive Normalization for Single Domain Generalization. CVPR 2021
- Tent: Fully Test-time Adaptation by Entropy Minimization. ICLR 2021
- Uncertainty Modeling for Out-of-Distribution Generalization. ICLR 2022
Variational Inference
Augmentation
Registration
- A Log-Euclidean Polyaffine Framework for Locally Rigid or Affine Registration. WBIR 2006
- A Fast Diffeomorphic Image Registration Algorithm. Neuroimage 2007
- Symmetric Diffeomorphic Image Registration with Cross-correlation: Evaluating Automated Labeling of Elderly and Neurodegenerative Brain. MedIA 2008
- An Unsupervised Learning Model for Deformable Medical Image Registration. CVPR 2018
- Unsupervised Learning for Fast Probabilistic Diffeomorphic Registration. MICCAI 2018
- Weakly-supervised Convolutional Neural Networks for Multi-modal Image Registration. MedIA 2018
- Conditional Deformable Image Registration with Convolutional Neural Network. MICCAI 2021
- SynthMorph: Learning Contrast-invariant Registration without Acquired Images. IEEE-TMI 2022
- Learning Conditional Deformable Templates with Convolutional Networks. NeurIPS 2019
- HyperMorph: Amortized Hyperparameter Learning for Image Registration. IPMI 2021
- Image-to-Graph Convolutional Network for 2D/3D Deformable Model Registration of Low-Contrast Organs. IEEE-TMI 2022
- Region-specific Diffeomorphic Metric Mapping. NeurIPS 2019
Reinforcement Learning
- Human-level control through deep reinforcement learning. Nature 2015
- Efficient Reinforcement Learning Through Trajectory Generation. Frontiers in Neurorobotics 2022
- Safe Model-Free Reinforcement Learning using Disturbance-Observer-Based Control Barrier Functions. Frontiers in Robotics and AI 2021
Long Term Dependency
Differential Equation
- Neural Ordinary Differential Equations. NeurIPS 2018
- Neural Flows: Efficient Alternative to Neural ODEs. NeurIPS 2021
Practical Machine Learning