| US 7,477,784 B2 | ||
| Spatial transforms from displayed codes | ||
| Jian Wang, Beijing (China); Liyon Chen, Beijing (China); Yihua Xu, Beijing (China); Yingnong Dang, Beijing (China); and Zhouchen Lin, Beijing (China) | ||
| Assigned to Microsoft Corporation, Redmond, Wash. (US) | ||
| Filed on Mar. 01, 2005, as Appl. No. 11/67,948. | ||
| Prior Publication US 2006/0204101 A1, Sep. 14, 2006 | ||
| This patent is subject to a terminal disclaimer. | ||
| Int. Cl. G06K 9/62 (2006.01); G06K 9/40 (2006.01) | ||
| U.S. Cl. 382—209 [382/254] | 20 Claims |
| 1. A process for processing an image containing codes comprising the steps of:
using a processor or computer to perform the steps of:
identifying code elements within the image;
based on the location of the identified code elements, determining candidates for main directions of parallel lines within
the image;
identifying pairs of near perpendicular directions from the candidates;
determining two main directions from the pairs of near perpendicular directions, the two main directions corresponding to
the directions of lines forming a grid associated with the identified code elements;
based on the two main directions, extracting features from the image;
generating a homography matrix from the extracted features and from the two main directions;
associating at least one of the identified code elements with one of the lines of the grid; and
refining the homography matrix by attempting to minimize the distance between the identified code elements and the lines of
the grid.
|