Digital-camera captured image restoration by using spline-based dewarping function and deshading function based on polynomial-logarithm

Abstract

This paper proposes a novel dewarping and deshading technique for digital camera based document images, from which images are warped on the curve surface of the book bound volume. Our dewarping technique by using look-up table transforms pixels on the curve coordinate system into the planar coordinate system. In the deshading method, the shade on the image is removed by using Least Square Error method (LSE). Our technique is independent on document contents, page layout, and font styles. In the experiments, the test document images composed various font styles and sizes printed in common text books. The validity of our methods is compared with the program’s curvature correction of Abby FineReader Professional 9.0 and Omnipage professional 16. Our technique outperforms the Abby’s curvature correction up to 49 percent and the OmniPage’s curvature correction up to 10 percent for word and character precisions.