Improving OCR for Historical Texts of Multiple Languages