SF-FSDA: Source-Free Few-Shot Domain Adaptive Object Detection with Efficient Labeled Data Factory.
Han SunRui GongKonrad SchindlerLuc Van GoolPublished in: CoRR (2023)
Keyphrases
- labeled data
- target domain
- domain adaptation
- transfer learning
- object detection
- unlabeled data
- semi supervised
- semi supervised learning
- source domain
- text classification
- training data
- cross domain
- active learning
- domain specific
- multiple sources
- supervised learning
- labeled examples
- data points
- semi supervised classification
- labeled training data
- prior knowledge
- class labels
- learning algorithm
- object categories
- data sets
- visual features
- training samples
- computer vision
- unsupervised learning
- labeling process
- labeled instances
- multi class
- information extraction
- training set
- object recognition
- machine learning
- real world