Generalization of deep learning models for ultra-low-count amyloid PET/MRI using transfer learning.
Kevin T ChenMatti SchürerJiahong OuyangMary Ellen I KoranGuido DavidzonElizabeth MorminoSolveig TiepoltKarl-Titus HoffmannOsama SabriGreg ZaharchukHenryk BarthelPublished in: European journal of nuclear medicine and molecular imaging (2020)
Deep learning can successfully produce diagnostic amyloid PET images from short frame reconstructions. Data bias should be considered when applying pre-trained deep ultra-low-count amyloid PET/MRI networks for generalization.
Keyphrases
- deep learning
- computed tomography
- pet ct
- positron emission tomography
- convolutional neural network
- contrast enhanced
- artificial intelligence
- magnetic resonance imaging
- pet imaging
- high resolution
- machine learning
- diffusion weighted imaging
- peripheral blood
- big data
- electronic health record
- magnetic resonance
- optical coherence tomography
- data analysis