An efficient semi-supervised quality control system trained using physics-based MRI-artefact generators and adversarial training.
Daniele RavìFrederik BarkhofDaniel C. AlexanderGeoffrey J. M. ParkerArman EshaghiPublished in: CoRR (2022)
Keyphrases
- control system
- semi supervised
- training process
- training set
- supervised learning
- high quality
- medical images
- semi supervised learning
- magnetic resonance imaging
- discriminative training
- fuzzy logic
- data sets
- multi layer perceptron
- high resolution
- supervised training
- training examples
- training algorithm
- multilayer neural network
- training procedure
- training stage
- semi supervised classification
- fully labeled
- magnetic resonance images
- deformable models
- labeled data
- image formation
- quality measures
- labeled examples
- control algorithm
- mr images
- backpropagation algorithm
- multi view
- echo state networks
- hidden markov models
- contrastive divergence
- pairwise