Spatial Consistency Loss for Training Multi-Label Classifiers from Single-Label Annotations.
Thomas VerelstPaul K. RubensteinMarcin EichnerTinne TuytelaarsMaxim BermanPublished in: CoRR (2022)
Keyphrases
- multi label
- multi label classification
- image annotation
- classifier training
- class labels
- multilabel classification
- image classification
- text categorization
- training examples
- label space
- graph cuts
- binary classification
- multi instance
- training set
- multiple labels
- text classification
- labeled images
- training samples
- multi label image annotation
- automatic image annotation
- supervised learning
- multi label learning
- text classifiers
- training data
- label assignment
- binary classifiers
- classification algorithm
- labeled data
- training dataset
- multi class
- max margin
- information retrieval
- semi supervised learning
- visual features
- decision trees
- feature selection
- naive bayes
- k nearest neighbor
- higher order
- nearest neighbor
- semi supervised
- knn
- support vector
- data sets