Solving jigsaw puzzles by computer
Annals of Operations Research
Corner detection using bending value
Pattern Recognition Letters
SUSAN—A New Approach to Low Level Image Processing
International Journal of Computer Vision
Digital Image Processing
Nonparametric Segmentation of Curves into Various Representations
IEEE Transactions on Pattern Analysis and Machine Intelligence
A shape and image merging technique to solve jigsaw puzzles
Pattern Recognition Letters
Document Reconstruction Based on Feature Matching
SIBGRAPI '05 Proceedings of the XVIII Brazilian Symposium on Computer Graphics and Image Processing
Globally Consistent Reconstruction of Ripped-Up Documents
IEEE Transactions on Pattern Analysis and Machine Intelligence
Hi-index | 0.00 |
The paper proposes a new algorithm for reconstruction of torn document ripped-up by hand using a unique feature, Edge Envelope. The essence of the paper is to match the edges using high curvature points and edge variations as human visual system does while reconstructing the document. Moreover the ambiguity created during the matching of edges due to shearing can be overcome by smoothing the edge envelopes using a filter. Further, a newly defined parameter Normalized Edge Envelope Difference can be used as an effective quantitative measure for matching the edges of torn pieces, which exploits property of human visual system. The effectiveness of the algorithm is demonstrated through preliminary results by taking limited amount of torn pieces.