The Importance of Metric Learning for Robotic Vision: Open Set Recognition and Active Learning.
Benjamin J. MeyerTom DrummondPublished in: ICRA (2019)
Keyphrases
- metric learning
- robotic vision
- active learning
- semi supervised
- distance metric learning
- distance metric
- recognition rate
- pairwise
- machine learning and pattern recognition
- semi supervised learning
- person re identification
- dimensionality reduction
- multi task
- labeled data
- distance function
- feature extraction
- object recognition
- pattern recognition
- feature space
- learning tasks
- supervised learning
- recognition algorithm
- subject to linear constraints
- machine learning
- mahalanobis metric
- transfer learning
- training examples
- unsupervised learning
- random sampling
- reinforcement learning
- computer vision
- euclidean distance
- image classification
- multi class
- selective sampling
- image features
- knowledge base