Source-Free Domain Adaptation for RGB-D Semantic Segmentation with Vision Transformers.
Giulia RizzoliDonald ShenajPietro ZanuttighPublished in: WACV (Workshops) (2024)
Keyphrases
- domain adaptation
- semantic segmentation
- multiple sources
- target domain
- superpixels
- cross domain
- conditional random fields
- scene classification
- labeled data
- semi supervised
- transfer learning
- object categories
- computer vision
- object segmentation
- domain specific
- sentiment classification
- test data
- training samples
- object detection
- training data
- co training
- semi supervised learning
- image processing
- active learning
- learning algorithm
- feature selection
- object recognition
- pairwise
- image representation
- unlabeled data