Title :
Trademark Image Retrieval Based on Shape and Key Local Color Features
Author :
Wang, Yong-jiao ; Zheng, Chun-feng
Author_Institution :
Dept. of Comput. Sci., Pingdingshan Inst. of Technol., Pingdingshan, China
Abstract :
An combined trademark image retrieval method based on shape and the key local color features is proposed. First , retrieve images with a wavelet modulus maxima algorithm based on the self-adaptive threshold using the shape features. Second, retrieve using the key local color features which contained the colorspsila statistic and local distributing characteristics. At last, set a weight considering the similarity degree of the shape and color. The experimental results indicated that this algorithm can retrieve similar trademark images fast and efficiently, and improve the recall ratio and precision ratio.
Keywords :
feature extraction; image colour analysis; image retrieval; image segmentation; shape recognition; statistical analysis; trademarks; wavelet transforms; key local color feature; self-adaptive threshold; shape feature; trademark image retrieval; wavelet modulus maxima algorithm; Computer science; Feature extraction; H infinity control; Image converters; Image retrieval; Information retrieval; Shape measurement; Statistical distributions; Trademarks; Transportation; key local color; self-adaptive threshod; similarity degree; trademark image retrieval; wavelet modulus maxima;
Conference_Titel :
Information and Computing Science, 2009. ICIC '09. Second International Conference on
Conference_Location :
Manchester
Print_ISBN :
978-0-7695-3634-7
DOI :
10.1109/ICIC.2009.193