Document perceptual quality ground truth creation

Perspective distortion always occurs while scanning thick, bound documents, resulting in two problems in the scanned grayscale image - (i) shade along the 'spine' of the book, and (ii) warping of words in the shade area. We proposed a restoration system to solve these two problems in our previous paper [1]. However the shape of the warped words was not fully restored, since we simply shifted and rotated the warped word. In this paper, we will propose a better method to straighten warped text line using polynomial regression. The implementation results comparing with those of the old system have shown a marked improvement in the image restoration.