| 摘要 |
<p>A 3D printer that uses two video cameras with different magnification powers for acquiring images from 2D negatives for editing and key subject alignment. One video camera takes full view image for key subject selection and picture cropping, the other video camera which is used for acquiring image data for key subject alignment sees only an enlarged section of the image around the key subject. The printer also uses an automated image matching algorithm to find relative location of the key subject in each frame. In particular, the algorithms are divided into a crude computation step and a fine computation step which reduces computation time. The printer uses a criterion to check whether the actual key subject alignment is satisfactory. The optics for printing and key subject alignment are assembled on a single mechanical stage to reduce mechanical errors.</p> |
