Localizing text in color images
Jin Zhao-hui · Journal of Circuits and Systems · 2006
With the advances of content-based retrieval technology,people have realized the importance of semantics for text embedded in color image and video text.Extracting and recognizing the texts in color images needs firstly to localizing them.A new connected component based segmentation algorithm which automatically extracts text regions from color images is proposed in this paper.A multi-channel decomposition method is utilized to get connected components,then connected component analysis and alignment analysis are used to localizing the texts in color images.The algorithm can detect not only horizontal text and vertical text but also declining text.Experimental results show the proposed algorithm can localize text effectively.