Research Document Imaging Hybrid Processing Previous Work

Search

Previous Work in Hybrid Processing
Dr. Eugene Ageenko

Numerous techniques and systems have been proposed for line-drawing images, to extract semantic information and perform a conversion of the document to vector space [Hou69, HK83, Ab89, Lea93, NL90, Kas90, RM95, KBO96, WD99]. However, no one solution creates a faithful copy of the original document, and expensive human interaction is often required [RM95].

The hybrid raster/vector compression system has been proposed to eliminate the necessity of converting the engineering drawings into vector format in order to process them in CAD applications [Wils96, 99]. Typically, vector and raster representations of the image are stored together in the same file. Existing information is kept in original raster form when new edits are made in vector form native for CAD. The vector features of the raster object can be extracted on demand, so that this raster object can be processed (moved, scaled, rotated or removed) directly in the raster form [SEA].

References.