Leveraging Uncertainty for Deep Interpretable Classification and Weakly-Supervised Segmentation of Histology Images.
Soufiane BelharbiJérôme RonyJose DolzIsmail Ben AyedLuke McCaffreyEric GrangerPublished in: CoRR (2022)
Keyphrases
- weakly supervised
- segmentation algorithm
- segmentation method
- image classification
- fully supervised
- test images
- superpixels
- input image
- image database
- image regions
- ground truth
- image segmentation
- semantic segmentation
- feature extraction
- object class
- image features
- bounding box
- semi supervised
- machine learning
- feature space
- energy function
- level set
- image retrieval
- multiscale
- relation extraction
- object detectors
- multiple objects
- supervised learning
- feature vectors
- feature set
- medical images
- object recognition
- region of interest
- text classification
- shape model
- conditional random fields
- spatial information
- viewpoint
- unsupervised learning