A semantic-based image retrieval system: VisEngine
Jian-Yong Sun, Zhengxing Sun, Ri Zhou, Huifeng Wang · 2003
A semantic-based image retrieval system is proposed in which, prior to retrieval, main regions with plenty of semantic information are segmented from an image. Then a semantic visual template is created for these regions on the basis of weighted feature combination. These templates are stored in XML files. During retrieval, three kinds of query patterns are supplied for the user and a mining algorithm is proposed to find out which feature the user has interest in mostly in feedback images and which makes semantic information users give used adequately by the computer. Experimental results show good performance of the system based on 3400 images.