Hidden stratification causes clinically meaningful failures in machine learning for medical imaging.
Luke Oakden-RaynerJared DunnmonGustavo CarneiroChristopher RéPublished in: CHIL (2020)
Keyphrases
- medical imaging
- machine learning
- medical images
- x ray
- image registration
- medical image processing
- medical image analysis
- anatomical structures
- shape analysis
- image analysis
- computer vision
- computed tomography
- imaging modalities
- magnetic resonance images
- ct images
- image processing
- image segmentation
- image processing and analysis
- magnetic resonance imaging
- biomedical engineering
- clinical practice
- image guided surgery
- computer aided diagnosis
- remote sensing
- computer vision and image processing
- image guided
- neural network
- pattern recognition
- volumetric data
- mri scans
- surface registration
- feature selection
- medical image segmentation
- artificial intelligence
- medical education
- three dimensional
- image restoration
- feature extraction
- computer assisted