# SealClub: Computer-aided Paper Document Authentication

arXiv: https://arxiv.org/abs/2210.07884 · HTML: https://mlsanity.com/paper/2210.07884/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 14 Oct 2022.

## Built on by (most cited first, 1 of 1)

- [Words as Geometric Features: Estimating Homography using Optical Character Recognition as Compressed Image Representation](https://mlsanity.com/paper/2505.18925/) (2025)

## Abstract (arXiv)

Digital authentication is a mature field, offering a range of solutions with rigorous mathematical guarantees. Nevertheless, paper documents, where cryptographic techniques are not directly applicable, are still widely utilized due to usability and legal reasons. We propose a novel approach to authenticating paper documents using smartphones by taking short videos of them. Our solution combines cryptographic and image comparison techniques to detect and highlight subtle semantic-changing attacks on rich documents, containing text and graphics, that could go unnoticed by humans. We rigorously analyze our approach, proving that it is secure against strong adversaries capable of compromising different system components. We also measure its accuracy empirically on a set of 128 videos of paper documents, half containing subtle forgeries. Our algorithm finds all forgeries accurately (no false alarms) after analyzing 5.13 frames on average (corresponding to 1.28 seconds of video). Highlighted regions are large enough to be visible to users, but small enough to precisely locate forgeries. Thus, our approach provides a promising way for users to authenticate paper documents using conventional smartphones under realistic conditions.