BINARIZATION TECHNIQUE FOR THE DEGRADED DOCUMENT IMAGES AND INSCRIPTION IMAGES

Abstract

Over a decade,text extraction of document has been a subject of intrest for the research,but a very few work has been done in digitizing inscription images of historical monuments. For unclear and complex archaeological inscription images,there is no sharp distinction between foreground and background.There are several problems in the text of inscription images,such problems are like there is low contrast between text and background thus the use of previously available method unsuitable. For the regions having high edge density and strength simple edge-based approaches are also considered useful. This edge-based method give good result if background is not complex,but for the inscription images background is complex,thus this method cannot be used directly. Badly degraded images which is having high inter/intravariation between the background and the foreground text,the segmentation of the text becomes a big challenge. This thesis propose a novel document image binarization technique for the monument inscription that earliar was used for the binarization of degrade document images.The proposed method is basically an adaptive image binariztion technique.In this first an adaptive contrast map is constructed for the input inscription image.The contrast map is then binarizd and then combined with canny’s edge map to identify the text stroke edge pixels.Local threshold that is estimated on the intensities of detected text stroke edge pixels within a local window,is further applied for the document tex segmentation.the proposed method is very simple, robust and it involves minimum parameter tuning. It has been applied on different monuments inscription images and have given good results.