Reverse Classification Accuracy: Predicting Segmentation Performance in the Absence of Ground Truth.
Vanya V. ValindriaIoannis LavdasWenjia BaiKonstantinos KamnitsasEric O. AboagyeAndrea G. RockallDaniel RueckertBen GlockerPublished in: CoRR (2017)
Keyphrases
- ground truth
- classification accuracy
- ground truth data
- image segmentation algorithms
- segmented images
- quantitative evaluation
- segmentation algorithm
- test images
- shape prior
- level set
- image segmentation
- segmentation method
- feature selection
- contour detection
- image analysis
- feature space
- fully automatic
- multiscale
- image processing
- segmentation accuracy
- grey level
- training data
- optical flow
- high quality
- manually labeled
- optimal segmentation
- data sets
- texture segmentation
- segmentation errors
- brain mri
- point to point correspondences
- gold standard
- support vector
- training set
- active contours
- markov random field
- object detection