Abstract: Automated digitization of government documents containing mixed script is a challenging task. Accurate segmentation of these documents boosts the accuracy of OCR systems. An accurately ...