SPL: Exploiting Unlabeled Data for Multi-label Image Classification.
Weibo ZhangFuqing ZhuJiao DaiSonglin HuJizhong HanTao GuoPublished in: ICME (2019)
Keyphrases
- multi label
- unlabeled data
- image classification
- labeled data
- semi supervised learning
- text categorization
- semi supervised
- multi label learning
- class labels
- text classification
- multi label classification
- active learning
- multiple labels
- supervised learning
- training data
- co training
- data points
- labeled training data
- feature extraction
- binary classification
- learning tasks
- image representation
- image annotation
- image features
- visual features
- multi instance
- learning algorithm
- training examples
- training set
- classifier training
- graph cuts
- machine learning
- label propagation
- class distribution
- scene classification
- prior knowledge
- feature selection
- image processing
- data analysis
- label information
- small number
- cost sensitive
- k nearest neighbor
- transfer learning
- unsupervised learning
- svm classifier
- naive bayes