Document image dewarping using robust estimation of curled text lines

摘要: Digital cameras have become almost ubiquitous and their use for fast casual capturing of natural images is unchallenged. For making documents, however, they not caught up to flatbed scanners yet, mainly because camera tend suffer from distortion due the perspective are therefore limited in further archival or OCR. non-planar paper surfaces like books, page curl causes additional distortion, which poses an even greater problem its nonlinearity. This presents a new algorithm removing both distortion. It requires only single image as input relies on priori layout information instead hardware. Therefore, it much more user friendly than most previous approaches, allows flexible ad hoc document capture. Results presented showing that produces visually pleasing output increases OCR accuracy, thus having potential general purpose preprocessing tool based