DocumentCode :
3190509
Title :
Study and application on automatic tagging algorithm of image semantic information
Author :
Liu, Zhihui ; Cao, Yan ; Mu, Xiangwei
Author_Institution :
Transp. Manage. Inst., Dalian Maritime Univ., Dalian, China
fYear :
2011
fDate :
8-10 Aug. 2011
Firstpage :
6767
Lastpage :
6770
Abstract :
Image annotation is an indispensable step in the process of CBIR (content-based image retrieval). It comprehensively considered both features of the image visual and text messages, which can improve the accuracy of the content-based image retrieval and make image search system more accurate when getting target image. Based on the study about the mainstream technology of the current image mark methods, using CMRM algorithm as machine learning model, this paper realized an image semantic automatic tagging module by combining training methods of image texture characteristics and image retrieval technologies based on color. The merger of this module and early results of CBIR enabled the combination of content-based retrieval and keyword retrieval. It made some improvements to the retrieval performance and narrowed the gap of semantics. Experimental results demonstrated that this project can to a certain extent help users more precisely retrieve to their target images more precise.
Keywords :
content-based retrieval; image colour analysis; image retrieval; image texture; learning (artificial intelligence); CMRM algorithm; automatic tagging algorithm; content-based image retrieval; image annotation; image color; image mark methods; image search system; image semantic information; image texture characteristics; image visual; keyword retrieval; machine learning model; mainstream technology; text messages; training methods; Accuracy; Feature extraction; Image color analysis; Image retrieval; Semantics; Tagging; Visualization; Auto-Image Annotation; CBIR; CMRM;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Artificial Intelligence, Management Science and Electronic Commerce (AIMSEC), 2011 2nd International Conference on
Conference_Location :
Deng Leng
Print_ISBN :
978-1-4577-0535-9
Type :
conf
DOI :
10.1109/AIMSEC.2011.6011470
Filename :
6011470
Link To Document :
بازگشت