Relevance feedback for CBIR: a new approach based on probabilistic feature weighting with positive and negative examples.
Mohammed Lamine KherfiDjemel ZiouPublished in: IEEE Trans. Image Process. (2006)
Keyphrases
- feature weighting
- relevance feedback
- positive and negative examples
- positive examples
- image retrieval
- training set
- active learning
- negative examples
- text categorization
- hypothesis space
- positive and negative
- information retrieval systems
- version space
- low level features
- document retrieval
- multiple features
- training examples
- query expansion
- text classification
- information retrieval
- feature selection
- tree patterns
- retrieval model
- logic programs
- tf idf
- visual features
- high level
- learning algorithm