Deep Learning With Anatomical Priors: Imitating Enhanced Autoencoders In Latent Space For Improved Pelvic Bone Segmentation In MRI.
Duc Duy PhamGurbandurdy DovletovSebastian WarwasStefan LandgräberMarcus JägerJosef PauliPublished in: ISBI (2019)
Keyphrases
- deep learning
- bone segmentation
- medical images
- ct images
- fracture detection
- clinical applications
- treatment planning
- computed tomography
- anatomical structures
- medical imaging
- imaging modalities
- ct scans
- unsupervised learning
- x ray images
- active shape model
- machine learning
- magnetic resonance
- mr images
- three dimensional
- prior information
- deformable models
- image guided
- prior knowledge
- latent variables
- region of interest
- high resolution
- data mining
- bayesian framework
- gaussian process
- denoising
- accurate segmentation
- feature space