Uncertainty-Based Method for Improving Poorly Labeled Segmentation Datasets.
Ekaterina RedekopAlexey ChernyavskiyPublished in: ISBI (2021)
Keyphrases
- segmentation method
- segmentation algorithm
- detection method
- cost function
- fully automatic
- significant improvement
- test images
- edge detection
- energy function
- image segmentation
- experimental results on real world
- piece wise
- thresholding method
- segmentation accuracy
- grey level
- synthetic data
- dynamic programming
- clustering method
- high accuracy
- supervised learning
- prior information
- support vector machine
- training data
- similarity measure
- neural network