MaskSplit: Self-supervised Meta-learning for Few-shot Semantic Segmentation.
Mustafa Sercan AmacAhmet SencanOrhun Bugra BaranNazli Ikizler-CinbisRamazan Gokberk CinbisPublished in: WACV (2022)
Keyphrases
- meta learning
- semantic segmentation
- inductive learning
- learning tasks
- conditional random fields
- superpixels
- model selection
- scene classification
- weakly supervised
- object categories
- decision trees
- machine learning algorithms
- video sequences
- data mining
- feature selection
- visual features
- object recognition
- base classifiers
- object classes
- machine learning
- metamodel
- natural images
- image set
- semi supervised
- probabilistic model
- multiscale
- computer vision
- learning algorithm