A stochastic optimization approach for shredded document reconstruction in forensic investigations

Area of Science:

• Forensic Science
• Computer Science
• Optimization Theory

Background:

• Document shredding is a common security measure, posing challenges for forensic document recovery.
• Traditional reconstruction methods like physical edge matching are inadequate for cross-cut shredded documents.

Purpose of the Study:

• To develop an advanced computational method for reconstructing shredded documents.
• To overcome limitations of existing methods in forensic investigations.

Main Methods:

• A stochastic optimization approach inspired by Markov chain Monte Carlo (MCMC) methods.
• Evaluation of visual content matches using edge compatibility metrics.
• Gamma distribution modeling of edge deviations with maximum likelihood estimation.

Main Results:

• The method successfully reconstructs diverse document types (typed, handwritten, photos).
• Outperforms simulated annealing (SA) and genetic algorithms (GA) in reconstruction success rates.
• Demonstrates practical utility on real shredded documents from the DARPA Shredder Challenge.

Conclusions:

• The developed MCMC-inspired method offers a robust and flexible framework for shredded document reconstruction.
• Advances forensic document examination capabilities, particularly for challenging cross-cut shredding.
• Semi-automated approach with human guidance enhances efficiency and accuracy for complex cases.