On Image Segmentation With Noisy Labels: Characterization and Volume Properties of the Optimal Solutions to Accuracy and Dice.
Marcus NordströmHenrik HultJonas SöderbergFredrik LöfmanPublished in: CoRR (2022)
Keyphrases
- image segmentation
- optimal solution
- high accuracy
- active contours
- multiscale
- computational cost
- boundary detection
- region growing
- computational efficiency
- segmentation method
- class labels
- prediction accuracy
- error rate
- level set method
- training set
- segmentation algorithm
- segmented images
- unseen data
- label noise
- total cost
- graph cuts
- markov random field
- image analysis
- lower bound
- training data