Classification d'images en apprenant sur des échantillons positifs et non labélisés avec un réseau antagoniste génératif(Image classification by learning on positive and unlabeled samples with a generative adversarial network).
Florent ChiaroniMohamed-Cherif RahalFrédéric DufauxNicolas HueberPublished in: CNIA+RJCIA (2018)
Keyphrases
- image classification
- unsupervised learning
- supervised learning
- learning process
- unlabeled samples
- incremental learning
- machine learning
- labeled and unlabeled data
- labeled samples
- active learning
- classification accuracy
- training set
- reinforcement learning
- feature extraction
- support vector
- feature vectors
- training samples
- unlabeled data
- data sets
- semi supervised classification
- decision trees
- object categories
- dimension reduction
- learning tasks
- generative model
- text classification
- semi supervised
- prior knowledge