Half a Percent of Labels is Enough: Efficient Animal Detection in UAV Imagery Using Deep CNNs and Active Learning.
Benjamin KellenbergerDiego MarcosSylvain LobryDevis TuiaPublished in: IEEE Trans. Geosci. Remote. Sens. (2019)
Keyphrases
- active learning
- aerial imagery
- object detection
- detection method
- training examples
- cellular neural networks
- detection accuracy
- anomaly detection
- image data
- training set
- path planning
- machine learning
- semi supervised
- data sets
- computationally expensive
- automatic detection
- e learning
- experimental design
- computer vision
- unmanned aerial vehicles