Mining High-Level User Concepts with Multiple Instance Learning and Relevance Feedback for Content-Based Image Retrieval.
Xin HuangShu-Ching ChenMei-Ling ShyuChengcui ZhangPublished in: Revised Papers from MDM/KDD and PAKDD/KDMCD (2002)
Keyphrases
- relevance feedback
- multiple instance learning
- high level
- low level features
- multiple instance
- image retrieval
- image annotation
- low level image features
- improve the retrieval accuracy
- multi class
- low level
- image categorization
- information retrieval systems
- active learning
- supervised learning
- image search
- query expansion
- retrieval process
- object based image retrieval
- information retrieval
- cbir systems
- concept learning
- semi supervised learning
- visual concepts
- semantic information
- metadata
- feature extraction