Two dimensional to three dimensional image conversion system design of digital archives for classical antiques and document

In this paper, we proposed an auto-generated two dimensional to three dimensional conversion system. We divided this architecture into two parts: In part one, we use Sobel edge detection [1] to segement the object for image of digital archives for classical antiques and document. Then, we used the Zhang and Suen thinning [2] to reduce the width of object's boundary to single pixel. After thinning, the image would be changed into cellular grid that likes a vein of leaf. In part two, we filled the corresponding depth values in the objects of the grid image, because the lower position of Chinese painting mostly looks nearer, and the higher one mostly looks farther. So we used the depth map from near to far. Finally, we used the χ-shaped scan path to fill the depth values in the grid image.