Book reader optimization using a time of flight imaging sensor

摘要

This study expands on our previous approach for correcting warping in images of a book spread, which in turn can be utilized to improve character recognition. The correction is made possible via depth map extraction using a time of flight sensor. Curvature corrections are achieved by modifying the lens equation to take into account different height points on the book spread. The corrections have been expanded not only to flatten the page, but to extend accordingly the characters in the image. In addition, current resolution limitation from the time of flight device is overcome by scaling and matching the pixel depth data to an image taken from a camera with higher resolution. The Implementation results support the assertion of improved reading accuracy, which in turn highlights the merits of using this approach based on the depth maps in order to correct for the book curvature.