Enhancing Recognition of Visual Concepts with Primitive Color Histograms via Non-sparse Multiple Kernel Learning.
Alexander BinderMotoaki KawanabePublished in: CLEF (2) (2009)
Keyphrases
- visual concept detection
- multiple kernel learning
- visual concepts
- color histogram
- object recognition
- linear combination
- object categories
- image database
- spatial information
- visual features
- semi supervised learning
- image content
- feature selection
- image features
- visual content
- learning tasks
- image retrieval
- feature extraction
- high level
- kernel function
- learning algorithm
- kernel methods
- video content
- image collections
- semantic gap
- image annotation
- active learning
- multi task
- higher level