DocumentCode :
2353236
Title :
On solving 2D and 3D puzzles using curve matching
Author :
Kong, Weixin ; Kimia, Benjamin B.
Author_Institution :
Brown Univ., Providence, RI, USA
Volume :
2
fYear :
2001
fDate :
2001
Abstract :
We approach the problem of 2D and 3D puzzle solving by matching the geometric features of puzzle pieces three at a time. First, we define an affinity measure for a pair of pieces in two stages, one based on a coarse-scale representation of curves and one based on a fine-scale elastic curve matching method. This re-examination of the top coarse-scale matches at the fine scale results in an optimal relative pose as well as a matching cost which is used as the affinity measure for a pair of pieces. Pairings with overlapping boundaries are impossible and are removed from further consideration, resulting in a set of top valid candidate pairs. Second, triples arising from generic junctions are formed from this rank-ordered list of pairs. The puzzle is solved by a recursive grouping of triples using a best-first search strategy, with backtracking in the case of overlapping pieces. We also generalize aspects of this approach to matching of 3D pieces. Specifically, ridges of 3D fragments scanned using a laser range finder are detected using a dynamic programming method. A pair of ridges are matched using a generalization of the 2D curve matching approach to space curves by using an energy solution involving curvature and torsion, which are computed using a novel robust numerical method. The reconstruction of map fragments and broken tiles using this method is illustrated.
Keywords :
backtracking; dynamic programming; edge detection; feature extraction; image matching; stereo image processing; 2D puzzle solving; 3D puzzle solving; affinity measure; backtracking; best-first search strategy; broken tile reconstruction; coarse-scale representation curves; curvature; dynamic programming method; energy solution; fine-scale elastic curve matching method; generic junctions; geometric feature matching; map fragment reconstruction; matching cost; optimal relative pose; overlapping boundaries; pairings; puzzle pieces; rank-ordered list pairs; recursive triple grouping; ridge detection; robust numerical method; space curves; torsion; Art; Computational efficiency; Cost function; Dynamic programming; Failure analysis; Lapping; Robustness; Search problems; Shape; Tiles;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computer Vision and Pattern Recognition, 2001. CVPR 2001. Proceedings of the 2001 IEEE Computer Society Conference on
ISSN :
1063-6919
Print_ISBN :
0-7695-1272-0
Type :
conf
DOI :
10.1109/CVPR.2001.991015
Filename :
991015
Link To Document :
بازگشت