Adversarial Semi-Supervised Domain Adaptation for Semantic Segmentation: A New Role for Labeled Target Samples.
Marwa KechaouMokhtar Z. AlayaRomain HéraultGilles GassoPublished in: CoRR (2023)
Keyphrases
- domain adaptation
- semi supervised
- unlabeled samples
- semantic segmentation
- weakly supervised
- labeled data
- target domain
- semi supervised learning
- labeled examples
- unlabeled data
- supervised learning
- labeled training data
- training samples
- conditional random fields
- pairwise
- training data
- co training
- transfer learning
- training examples
- object categories
- active learning
- superpixels
- generative model
- unsupervised learning
- training set
- multi view
- labeled and unlabeled data
- test data
- cross domain
- image set
- object class
- data mining
- reinforcement learning
- face recognition
- learning algorithm
- object classes
- class labels
- data sets
- text categorization
- d objects
- small number
- domain specific
- object detection