Anatomical context protects deep learning from adversarial perturbations in medical imaging.
Yi LiHuahong ZhangCamilo BermudezYifan ChenBennett A. LandmanYevgeniy VorobeychikPublished in: Neurocomputing (2020)
Keyphrases
- medical imaging
- deep learning
- medical images
- anatomical structures
- medical image processing
- x ray
- image analysis
- computed tomography
- medical image analysis
- ct images
- imaging modalities
- image processing and analysis
- magnetic resonance imaging
- magnetic resonance images
- image registration
- shape analysis
- brain structures
- unsupervised feature learning
- computer vision
- image segmentation
- image processing
- image guided surgery
- mr images
- mental models
- machine learning
- deformable models
- remote sensing
- unsupervised learning
- weakly supervised
- three dimensional
- object detection
- volumetric data
- reinforcement learning