The systematic profiling of false identity documents: method validation and performance evaluation using seizures known to originate from common and different sources

Simon Baechler,Vincent Terrasse,Jean-Philippe Pujol,Thibaud Fritz,Olivier Ribaux,Pierre Margot

DOI: https://doi.org/10.1016/j.forsciint.2013.07.022

2013-10-10

Abstract:False identity documents constitute a potential powerful source of forensic intelligence because they are essential elements of transnational crime and provide cover for organized crime. In previous work, a systematic profiling method using false documents' visual features has been built within a forensic intelligence model. In the current study, the comparison process and metrics lying at the heart of this profiling method are described and evaluated. This evaluation takes advantage of 347 false identity documents of four different types seized in two countries whose sources were known to be common or different (following police investigations and dismantling of counterfeit factories). Intra-source and inter-sources variations were evaluated through the computation of more than 7500 similarity scores. The profiling method could thus be validated and its performance assessed using two complementary approaches to measuring type I and type II error rates: a binary classification and the computation of likelihood ratios. Very low error rates were measured across the four document types, demonstrating the validity and robustness of the method to link documents to a common source or to differentiate them. These results pave the way for an operational implementation of a systematic profiling process integrated in a developed forensic intelligence model.