Adaptive fuzzy text segmentation in images with complex backgrounds using color and texture

摘要: Textual information present in images can help to achieve the aim of automatic content based annotation and retrieval images. In this paper, we address problem text segmentation (TS) with complex background for recognition purposes. The proposed TS method takes as input localized proceeds follows: First, number initial clusters is determined by analyzing colors image. Second, image pixels are clustered using defined first step. compactness evaluated each step improved iteratively avoid possible oversegmentation characters. Finally, an algorithm on a rating scheme determine cluster where classified. basis results instead visual results. Comparative experimental test set 2684 characters reported.