MUS-CDB: Mixed Uncertainty Sampling with Class Distribution Balancing for Active Annotation in Aerial Object Detection.
Dong LiangJing-Wei ZhangYing-Peng TangSheng-Jun HuangPublished in: CoRR (2022)
Keyphrases
- class distribution
- cost sensitive
- active learning
- class imbalance
- misclassification costs
- training set
- multi class
- unlabeled data
- training examples
- training data
- binary classification
- training samples
- test set
- class labels
- base classifiers
- random sampling
- data mining
- test data
- naive bayes
- semi supervised learning
- supervised learning
- semi supervised
- data sets
- probability estimates