Automatic and online detection of document falsification (M/F)
@French National Centre for Scientific Research posted 5 hours agoJob Description
General information
Job title: Automatic and online detection of document falsification (M/F)
Reference: UMR9010-VERONIQUE-025
Number of positions: 1
Work location: GIF SUR YVETTE
Publication date: Friday, June 6, 2025
Type of contract: Fixed-term contract Doctoral student
Contract duration: 36 months
Thesis start date: October 1, 2025
Workload: Full
Remuneration: The remuneration is a minimum of €2,200.00 per month
CN section(s): 41 – Mathematics and interactions of mathematics
Description of the thesis subject
Documents play a vital role in communication, record keeping, and decision-making across diverse domains, but their integrity is increasingly threatened by sophisticated digital manipulation techniques. Traditional forensic approaches rely on device-specific artifact analysis and content examination to verify document authenticity and detect forgeries. However, these methods often lack statistical rigor, leading to uncertainty about the reliability of detections. This project aims to advance document forensic analysis by
developing detection methods, a contrario, based on large deviation theory
. By modeling the document image formation process and analyzing intrinsic structures such as text line alignment and redundancy, we aim to establish a rigorous framework for anomaly detection.
Our approach will quantify the reliability of detections, thereby reducing false positives and improving decision-making. However, the challenge of this thesis is not limited to the design of the theory and algorithms, but also includes their publication online in an open form, so that anyone can submit a suspicious document and obtain an automatic scientific report on the detected anomalies.
Work context
The project will be carried out at the Borelli Center UMR9010 at ENS Paris-Saclay. Collaborations are planned with the “Laboratorio Del Plata” in Uruguay, where at least one month per year will be spent.
Constraints and risks
WITHOUT
