DocumentCode :
3087162
Title :
Evaluation of Relative Focus Map Based Image Indexing
Author :
Kovács, Levente ; Szirányi, Tamás
Author_Institution :
Hungarian Acad. of Sci., Budapest
fYear :
2007
fDate :
25-27 June 2007
Firstpage :
186
Lastpage :
191
Abstract :
This paper deals with performance evaluation of image retrieval based on our previously introduced blind deconvolution-based automatic relative focus map estimation method. After a short introduction into the focus extraction approach and its application for image indexing, we will concentrate on evaluating the effectiveness of the retrieval, also compared to sharpness-based retrieval using traditional edge or frequency information. We will show that relative focus maps can be efficiently used for retrieving similarly focused images and thus it is a viable addition to traditional image indexing features.
Keywords :
deconvolution; feature extraction; image retrieval; automatic relative focus map estimation; blind deconvolution; focus extraction; image indexing; image retrieval; sharpness-based retrieval; Bayesian methods; Data mining; Deconvolution; Focusing; Frequency estimation; Image reconstruction; Image retrieval; Indexing; Information retrieval; Layout;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Content-Based Multimedia Indexing, 2007. CBMI '07. International Workshop on
Conference_Location :
Bordeaux
Print_ISBN :
1-4244-1011-8
Electronic_ISBN :
1-4244-1011-8
Type :
conf
DOI :
10.1109/CBMI.2007.385410
Filename :
4275073
Link To Document :
بازگشت