On Image Segmentation With Noisy Labels: Characterization and Volume Properties of the Optimal Solutions to Accuracy and Dice.
Marcus NordströmHenrik HultFredrik LöfmanJonas SöderbergPublished in: NeurIPS (2022)
Keyphrases
- image segmentation
- optimal solution
- graph cuts
- high accuracy
- markov random field
- active contours
- computational efficiency
- deformable models
- metaheuristic
- prediction accuracy
- error rate
- image processing
- np hard
- search space
- pairwise
- training set
- multiscale
- label noise
- medical images
- contour detection
- noisy environments
- boundary detection
- level set method
- noisy data
- medical imaging
- training examples
- gray level
- computational cost
- supervised learning