Transfer Learning from Synthetic Data Applied to Soil-Root Segmentation in X-Ray Tomography Images.
Clément DouarreRichard SchieleinCarole FrindelStefan GerthDavid RousseauPublished in: J. Imaging (2018)
Keyphrases
- synthetic data
- x ray
- real image data
- transfer learning
- poor contrast
- digital x ray images
- tomographic images
- medical imaging
- three dimensional
- electron microscope
- computer tomography
- ct and mr images
- electron microscopy
- transmission electron microscopy
- dual energy
- low dose
- segmentation algorithm
- segmentation method
- projection images
- x ray images
- automated segmentation
- ct volume
- intraoperative
- data sets
- fully automatic
- real world
- edge detection
- ct scans
- fluoroscopic images
- mri data
- input image
- machine learning
- image segmentation
- object recognition
- labeled data
- image registration
- active learning
- reinforcement learning
- database
- automatic segmentation
- learning algorithm
- similarity measure
- noise level
- semi supervised learning
- volumetric data
- medical images
- accurate segmentation