Adaptive thresholding to robust image binarization for degraded document images

Abstract

Owing to the elevated intra/inter variation among the foreground and background text of various document images, the text segmentation from the poorly degraded document images is the difficult job. This paper presents the document image binarization method by adaptive image contrast which is the integration of the local image gradient and the local image contrast which is lenient to background and text variation generated by various document degradations. Initially, an adaptive contrast map is constructed for the input degraded document image by the proposed document image binarization method. Then, the binarization is performed on this adaptive contrast map and the binarized contrast map is integrated with the Canny's edge map for determining the text stroke edge pixels. After that, depends on the local threshold which is defined by the identified text stroke edge pixels' intensities in the local window, the document text is divided. The proposed method is straight forward, vigorous, and it requires least amount of parameter tuning. The experimentation is performed on DIBCO 2011 dataset and the results of the experimentation show that the proposed method achieved high performance than the state-of-the-art methods.