Effect of observer variability and training cases on U-Net segmentation performance.
Jordan D. FuhrmanPeter HalloranRowena YipArtit C. JirapatnakulClaudia I. HenschkeDavid F. YankelevitzMaryellen L. GigerPublished in: Medical Imaging: Image Perception, Observer Performance, and Technology Assessment (2020)
Keyphrases
- moving observer
- image segmentation
- multiscale
- training set
- object segmentation
- segmentation method
- training phase
- shape prior
- segmentation algorithm
- energy function
- image analysis
- fully automatic
- region growing
- training process
- segmentation accuracy
- training sessions
- test images
- neural network
- decision trees
- case base
- motion model
- motion segmentation
- medical imaging
- multiple objects
- x ray
- training algorithm
- medical images
- edge detection
- segmentation errors
- brain mri
- software product line
- supervised learning