Image text extraction based on morphology and color layering

Abstract

In order to extract text regions in the color image, this paper proposes an image text extraction method based on morphology and color layering. Firstly, we extract the edges with Sobel operator, and then extract rectangular regions and non-rectangular regions according to text features by using morphological methods. Finally, we handle these two types of regions respectively where in Color Layering Algorithm is implemented when dealing with non-rectangular regions. This paper also proposes Large Characters Repair Algorithm so that the method can also be applied to the images with texts of different fonts. During the research, we have found that Color Layering Algorithm can separate text from complex background effectively, which makes background removed more easily. The experimental results show that the proposed method has a high accuracy rate. © 2013 Springer Science+Business Media New York.