Mosaicing of Torn Document Images

Recovery of ripped-up documents is a problem that often arises in archival study and investigation science. The growing interest in featureless image mosaicing is because of its demand as it is still in incubation stage. The use of mosaicing torn document images is that it has reliable performance, fast document recovery capability and onscreen visual verification. Further, Most document reconstruction problems can be solved in two steps first, finding an initial set of matching fragment pairs, then resolving the ambiguity among these fragments to reconstruct the original document. The principle technique used here is the featureless image mosaicing technique. The proposed architecture for mosaicing of torn document images involves various stages. The first stage includes the scanning of the two fragments of the ripped up document such that the non-uniform sides face each other. The fragments include both uniform and non-uniform boundaries. We have also considered the non-uniform boundary values since, they had to be separated from the uniform boundary values. In present work, we have planned to design an algorithm to re-construct or mosaic a ripped-up document. We have corroborated the methodology by considering two fragments of the same torn/ripped up document.