An efficient new approach to archiving scanned documents that makes the text searchable and adds a watermark to images for copy protection and validation is reported in the inaugural issue of the International Journal of Reasoning-based Intelligent Systems.
The promise of the paperless office was never fulfilled. Indeed, as a society we seem to use more paper documents than at any time in history. However, there is now the additional demands of requiring paper documents to be archived, to be searchable, and to leave a virtual paper-trail for documentary providence.
Roumen Kountchev of the Technical University of Sofia working with Vladimir Todorov and Roumiana Kountcheva of T&K Engineering Co. in Sofia, Bulgaria have a new approach for efficient archiving and content protection of scanned documents, comprising texts and pictures.
Their approach takes the raw image of a scanned document and processes the text and images in the document in two different ways. The images are compressed using an algorithm known as IDP (inverse difference pyramid). This is used in a lossless manner for texts, graphics, medical or biological images, biometric visual information, such as fingerprints and signatures. For photographs where some loss is often acceptable the images can be compressed further using "lossy" IDP. The text is compressed in a lossless way that allows optical character recognition (OCR) to be carried out.
The IDP algorithm has the advantage of allowing the addition to the whole document of an extra embedded and invisible layer that contains a digital watermark. The watermark is password protected and can only be seen or removed by verified users, thus providing legal providence for an authenticated scanned document. The team points out that the watermark layer does not interfere with the scanned image or the text, whether legal document, fingerprint, or photograph.
However, if someone attempts to change the image, the invisible watermark will also be affected. Damage to the watermark will thus prove unauthorised editing of the document.
Cite This Page: