US 7,430,497 B2
Statistical model for global localization
Jian Wang, Beijing (China); Zheng Chen, Beijing (China); Xiaoxu Ma, Beijing (China); Yingnong Dang, Beijing (China); Liyong Chen, Beijing (China); and Jiang Wu, Beijing (China)
Assigned to Microsoft Corporation, Redmond, Wash. (US)
Filed on Oct. 31, 2002, as Appl. No. 10/284,421.
Prior Publication US 2004/0085302 A1, May 06, 2004
Int. Cl. G06F 17/00 (2006.01)
U.S. Cl. 703—2  [382/173; 382/177; 382/179; 382/187; 382/188; 382/189; 382/314; 382/316] 19 Claims
OG exemplary drawing
 
1. A process for determining a path of a pen having a camera capturing information, which is used to determine candidates for said pen's location with respect to a surface, comprising the steps of:
segmenting a surface into blocks;
analyzing the segmented surface to determine a relationship between the blocks of the segmented surface;
receiving a list of candidates, representative of locations of the pen with respect to the surface, for each image captured by said camera, the list of candidates being constrained by the determined relationship between the blocks of the segmented surface; and
determining a most likely path of said pen with respect to the surface based on a minimal movement from a previous image captured by said camera and at least one candidate of at least one received list of candidates,
wherein said process skips at least one candidate of a list of candidates for an image in determining said most likely path.