MRI-SegFlow: a novel unsupervised deep learning pipeline enabling accurate vertebral segmentation of MRI images.
Xihe KuangJason Pui Yin CheungHonghan WuSocrates DokosTeng ZhangPublished in: EMBC (2020)
Keyphrases
- mri images
- deep learning
- magnetic resonance imaging
- unsupervised learning
- medical images
- deep architectures
- mri brain images
- magnetic resonance images
- imaging modalities
- mri data
- weakly supervised
- brain tumors
- machine learning
- medical imaging
- magnetic resonance
- supervised learning
- mr images
- brain tissue
- low contrast
- medical image analysis
- computer vision
- multiscale
- semi supervised
- fully automatic
- model selection
- mental models
- image registration
- white matter
- high resolution
- expectation maximization
- image processing
- decision making
- deformable models
- data mining