Digital Jigsaw Puzzle Solving

Research Overview

  • We aim to teach computer to solve digital jigsaw puzzles that are arbitrarily cut.
  • Automatically solving jigsaw puzzles is a classic AI problem that has broad applications in forensics, archaeology, archival research, etc.
  • We study the reassembly of 2D image, 2D text document, and 3D geometric fragments.
  • The main technical challenges include that:

    (1) locally, correlation between adjacent fragments is weak, so obtaining effective fragment matching is difficult, and

    (2) globally, from unreliable local alignmetns, searching for globally optimal composition is highly expensive and has many incorrect local optimal solutions.


Reassembling Shredded Document Stripes Using Word-path Metric and Greedy Composition Optimal Matching Solver.

Y. Liang and X. Li

IEEE Transactions on Multimedia (TMM), 22(5):1168-1181, 2019.

[Bib] [Codes] [Demo]

JigsawNet: Shredded Image Reassembly Using Convolutional Neural Network and Loop-based Composition.

C. Le and X. Li

IEEE Transactions on Image Processing (TIP), 28(8):4000-4015, 2019.

[Bib] [Paper] [Codes]

Hierarchical Fragmented Image Reassembly using a Bundle-of-Superpixel Representation.

X. Li, K. Xie, W. Hong, and C. Liu

Computer Aided Geometric Design (CAGD), 71:220-230, 2019.

[Bib] [Demo]

3D Fragment Reassembly using Integrated Template Guidance and Fracture-Region Matching.

K. Zhang, M. Manhein, W. Waggenspack, and X. Li

International Conference on Computer Vision (ICCV), pp. 2138-2146, 2015.

[Bib] [Demo]

A Graph-based Optimization Algorithm for Fragmented Image Reassembly.

K. Zhang and X. Li

Graphical Models, Vol. 76, Issue 5, pp. 484-498, 2014.

[Bib] [Demo]

Locations of visitors to this page
Web Page Hit Counter