作者: M. Banerjee , M. K. Kundu
DOI: 10.1007/978-3-540-77046-6_18
关键词: Visual Word 、 Artificial intelligence 、 Set (abstract data type) 、 Pattern recognition 、 Content-based image retrieval 、 Selection (linguistics) 、 Image retrieval 、 Fuzzy logic 、 Weighting 、 Computer science 、 Relevance feedback
摘要: Content-Based Image retrieval has emerged as one of the most active research directions in past few years. In CBIR, selection desired images from a collection is made by measuring similarities between extracted features. It hard to determine suitable weighting factors various features for optimal when multiple are used. this paper, we propose relevance feedback frame work, which evaluates features, fuzzy entropy based feature evaluation index (FEI) considering both relevant well irrelevant set retrieved marked users. The results obtained using our algorithm have been compared with agreed upon standards visual content descriptors MPEG-7 core experiments.