With the popularity of cheap and portable cameras, camera-based document image processing and analysis becomes a very hot topic of high research values and broad application prospects in recent years. On the basis of the current progress of our national youth science fund project, we intend to study the correction and restoration of document images captured by hand-held cameras. These images are often corrupted due to many factors, such as camera perspective, surface curling, camera shaking, defocus and insufficient resolutions. Our study attempts to exploit the nature of these problems and establish a related basic theory under some reality constraints and reasonable assumptions, so as to propose a practicable solution and give rise to a set of key technologies on these challenging issues. Finally, effective algorithms will also be designed to enhance and restore the captured low-quality textual images. The achievements of this project will benefit many applications, including the new generation of high quality document images capture, online text analysis of intelligent mobile terminals and textual evidence collection in video surveillance.
