Causal importance of low-level feature selectivity for generalization in image recognition.
Jumpei UkitaPublished in: Neural Networks (2020)
Keyphrases
- image recognition
- low level features
- image retrieval
- image classification
- low level
- semantic gap
- visual features
- pattern recognition
- high level
- image representation
- face recognition
- higher level
- semantic information
- relevance feedback
- causal relationships
- image annotation
- feature representations
- image processing
- image database
- semantic concepts
- multiscale
- image content
- cbir systems
- databases
- correlation filters
- object recognition
- bayesian networks
- decision trees
- neural network