MMCL: Boosting Deformable DETR-Based Detectors with Multi-Class Min-Margin Contrastive Learning for Superior Prohibited Item Detection.
Mingyuan LiTong JiaHui LuBowen MaHao WangDongyue ChenPublished in: CoRR (2024)
Keyphrases
- multi class
- object detection
- multi class object detection
- learning algorithm
- boosting algorithms
- multi class classifier
- learning process
- base classifiers
- object detectors
- cost sensitive
- feature selection
- multiclass classification
- multi class classification
- binary classification
- support vector
- active learning
- support vector machine
- generalization error
- learning tasks
- large margin classifiers
- object recognition
- linear classifiers
- training set
- binary classifiers
- binary classification problems
- multi class boosting
- similarity measure