Unshredding of Shredded Documents: Computational Framework and Implementation

TL;DR

This paper introduces an optimal algorithm for reconstructing shredded documents, demonstrating its effectiveness across various document types including handwritten, typed, and images.

Contribution

It presents a novel optimal algorithm with quadratic complexity for unshredding documents, improving reconstruction accuracy and efficiency.

Findings

Effective reconstruction of handwritten, typed, and image documents

Algorithm achieves optimal quadratic time complexity

Demonstrated success across multiple document types

Abstract

A shredded document $D$ is a document whose pages have been cut into strips for the purpose of destroying private, confidential, or sensitive information $I$ contained in $D$. Shredding has become a standard means of government organizations, businesses, and private individuals to destroy archival records that have been officially classified for disposal. It can also be used to destroy documentary evidence of wrongdoings by entities who are trying to hide $I$. In this paper, we present an optimal $O((n\times m)^2)$ algorithm $A$ that reconstructs an $n$-page $D$, where each page $p$ is shredded into $m$ strips. We also present the efficacy of $A$ in reconstructing three document types: hand-written, machine typed-set, and images.