INVESTIGATION OF INVARIANT COLOR MODELS FOR TEXT DETECTION FROM SCENERY IMAGES

Accession number;05A0212965
Title;INVESTIGATION OF INVARIANT COLOR MODELS FOR TEXT DETECTION FROM SCENERY IMAGES
Author; YANG X (Tokyo Inst. Technol., Tokyo, Jpn) TAKAHASHI H (Tokyo Inst. Technol., Tokyo, Jpn) NAKAJIMA M (Tokyo Inst. Technol., Tokyo, Jpn)
Journal Title;IEIC Technical Report (Institute of Electronics, Information and Communication Engineers)
Journal Code:S0532B
ISSN:0913-5685
VOL.104;NO.544(IE2004 124-144);PAGE.91-94(2005)
Figure&Table&Reference;FIG.6, REF.10
Pub. Country;Japan
Language;English
Abstract;This paper introduces an investigation of the color information to get good representation of text part from natural scenes. Two different categories are discussed based on the experiment results. Edge detection for text images is briefly reviewed to show a robust method for detecting the text edges in natural scenes, which is against the effects caused by shadow or highlight. The other is to try to find some color models, which can get good representation on the text part by applying a series of morphological operations under different image transitions. Many color models including the invariant color models proposed by us are tested to find the robust ones under arbitrary image conditions. For text detection tasks, both the clean representation and the intact edge are useful information. (author abst.)