Document Image Binarization Based on Combination of Global and Local Thresholding Methods

Binarization is an important pre-processing step in document image analysis. Binarization algorithms have been classified into two general categories of global threshold and local threshold methods, each of which has weaknesses and strengths. In this paper, to use the advantages of both methods, a mixed method is suggested for binarization. The proposed method is based on linear combination of Otsu's Global thresholding method and Niblack's local thresholding method. The proposed method has proper performance in terms of both global and local binarization in images with large smear and low contrast. The DBICO standard database is used to assess the performance of the proposed method. The efficiency of the proposed method using MPM, F and PSNR evaluation criteria has been compared with the classic binarization methods and the results of the experiments confirm the effectiveness of the proposed method in comparison with others.