Discrepancy-based Active Learning for Weakly Supervised Bleeding Segmentation in Wireless Capsule Endoscopy Images.
Fan BaiXiaohan XingYutian ShenHan MaMax Q.-H. MengPublished in: CoRR (2023)
Keyphrases
- capsule endoscopy
- weakly supervised
- active learning
- segmentation algorithm
- segmentation method
- narrow band
- image analysis
- computer aided
- image database
- tumor detection
- input image
- test images
- superpixels
- image data
- semantic segmentation
- ground truth
- image regions
- image features
- three dimensional
- semi supervised
- image segmentation
- level set
- multiscale
- machine learning
- bounding box
- multiple objects
- multiple images
- topic models
- image retrieval
- image classification
- medical images
- object segmentation
- region of interest
- object recognition
- image content
- training set
- object detectors
- denoising