Research on the Relevance Feedback-based Image Retrieval in Digital Library

In recent years, the relevance feedback technology is regarded in content-based image retrieval. This paper suggests a neural networks feedback algorithm based on the radial basis function, coming to extract the semantic character of image. The results of experiment indicated that the performance of this relevance feedback is better than the feedback algorithm based on Single-RBF.

• Rongtao Ding
• Xinhao Ji
• Linting Zhu

• Image retrieval
• relevance feedback
• radial basis function.

[1] Chen Wan- mi, Xie Zhi- bing, Fei Min- rui.Research on majorization
strategies of vision system of RoboCup.IEEE, 2004: 372- 375.
[2] Feng Jing, Mingjing Li, Hong-Jiang Zhang, Bo Zhang. Support Vector
Machines For Region-based Image Retrieval. Multimedia and Expo,
2003. ICME '03. Proceedings. 2003 International Conference on
,Volume: 2,6-9 July 2003: 21-24
[3] Paisarn Muneesawang,Ling Guan.An Interative Approach for CBIR
Using a Network of Radial Basis Function,IEEE Trans on
Multimedia.6(5),Oct.2004
[4] Tao Da- cheng, Tang Xiao- ou, Li Xue- long, et al.Asymmetric bagging
and random subspace for Support Vector Machines-based relevance
feedback in image retrieval J.IEEE Transaction on Pattern Analysis and
Machine Intelligence, 2006, 28(7).