Transfer metric learning for action similarity using high-level semantics.
Ziad Al-HalahLukas RybokRainer StiefelhagenPublished in: Pattern Recognit. Lett. (2016)
Keyphrases
- metric learning
- high level semantics
- distance metric
- distance function
- low level features
- euclidean distance
- learning tasks
- semi supervised
- pairwise
- low level
- distance measure
- high level
- multi task
- transfer learning
- similarity measure
- semantic gap
- image content
- data points
- dimensionality reduction
- feature space
- semi supervised learning
- visual features
- labeled data
- contextual information
- similarity search
- low level visual features
- visual information
- higher level
- semantic similarity
- machine learning
- markov random field
- supervised learning
- relevance feedback
- knn
- image data
- active learning
- image retrieval
- object recognition
- e learning
- computer vision