Solving jigsaw puzzles using image features

Authors:
Ture R. Nielsen;Peter Drewsen;Klaus Hansen
Affiliations:
Department of Computer Science, University of Copenhagen, Universitetsparken 1, DK-2100 Copenhagen, Denmark;Department of Computer Science, University of Copenhagen, Universitetsparken 1, DK-2100 Copenhagen, Denmark;Department of Computer Science, University of Copenhagen, Universitetsparken 1, DK-2100 Copenhagen, Denmark
Venue:
Pattern Recognition Letters
Year:
2008

Citing 4
Cited 4

Solving jigsaw puzzles by computer

Annals of Operations Research
A Statistical Approach to Texture Description of Medical Images: A Preliminary Study

CBMS '02 Proceedings of the 15th IEEE Symposium on Computer-Based Medical Systems (CBMS'02)
A shape and image merging technique to solve jigsaw puzzles

Pattern Recognition Letters
A global approach to automatic solution of jigsaw puzzles

Computational Geometry: Theory and Applications - Special issue on the 18th annual symposium on computational geometry—SoCG2002

Document analysis applied to fragments: feature set for the reconstruction of torn documents

DAS '10 Proceedings of the 9th IAPR International Workshop on Document Analysis Systems
A new technique for solving puzzles

IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics - Special issue on game theory
Analysis of document snippets as a basis for reconstruction

VAST'09 Proceedings of the 10th International conference on Virtual Reality, Archaeology and Cultural Heritage
Global consistency in the automatic assembly of fragmented artefacts

VAST'11 Proceedings of the 12th International conference on Virtual Reality, Archaeology and Cultural Heritage

Quantified Score

Hi-index	0.10

Visualization

Abstract

In this article, we describe a method for automatic solving of the jigsaw puzzle problem based on using image features instead of the shape of the pieces. The image features are used for obtaining an accurate measure for edge similarity to be used in a new edge matching algorithm. The algorithm is used in a general puzzle solving method which is based on a greedy algorithm previously proved successful. We have been able to solve computer generated puzzles of 320 pieces as well as a real puzzle of 54 pieces by exclusively using image information. Additionally, we investigate a new scalable algorithm which exploits the divide and conquer paradigm to reduce the combinatorially complex problem by classifying the puzzle pieces and comparing pieces drawn from the same group. The paper includes a brief preliminary investigation of some image features used in the classification.