LNPL-MIL: Learning from Noisy Pseudo Labels for Promoting Multiple Instance Learning in Whole Slide Image.
Zhuchen ShaoYifeng WangYang ChenHao BianShaohui LiuHaoqian WangYongbing ZhangPublished in: ICCV (2023)
Keyphrases
- multiple instance learning
- image categorization
- supervised learning
- multiple instance
- diverse density
- multi class
- image features
- learning problems
- learning algorithm
- unsupervised learning
- image data
- learning process
- image segmentation
- prior knowledge
- semi supervised
- image classification
- semi supervised learning
- active learning
- object recognition
- image collections
- image retrieval
- learning tasks
- image content
- image representation
- image analysis
- image regions
- real valued
- image annotation
- concept learning
- similarity measure
- low level
- positive bags
- machine learning
- localized content based image retrieval