This article reviews practical challenges to and perspectives on the use o f face recognition in criminal science. A better understanding of the limitations and opportunities of this technology is essential to determining how it should be used in the judicial system. We suggest that this reflection must be part of an interdisciplinary approach that integrates the use of algorithms while taking into account the unique characteristics of both criminology and forensic science. Determining the state of the art in the use of face recognition requires answering several fundamental questions. What sort of images are used in face recognition? How are they compared? What objectives does face recognition serve? Where is it implemented and by whom? These questions - and their answers - allow us to identify the challenges and shortcomings of face recognition, as well as perspectives on its development and research. At the data level, the main issue concerns the quality of the original image and its potential degradation during collection and storage, which can influence its subsequent use. At the methodological level, the stakes involve the lack of standardisation and transparency in tasks carried out by both humans and automatic systems. When looking at face recognition as used in the civil and judicial domains, questions arise around how to protect privacy and individual liberties. Finally, the main challenges raised by its use as evidence in court concern communication, as well as standardisation and methodological validation.

L’objectif de cet article est de cristalliser de manière pragmatique les enjeux et perspectives de la reconnaissance faciale en sciences criminelles pour acquérir une meilleure connaissance des limitations et des opportunités offertes par ces technologies, essentielles à leur application au sein du système judiciaire. Nous postulons que cette réflexion doit s’inscrire dans une approche interdisciplinaire qui intègre l’utilisation d’algorithmes en considérant les spécificités de la criminologie et de la science forensique. Cet éclairage permet d’établir un bilan de l’utilisation de la reconnaissance faciale en le scindant en questions fondamentales. Quelles sont les images utilisées en reconnaissance faciale ? Comment sont-elles comparées ? Quels objectifs la reconnaissance faciale sert-elle ? Où est-elle mise en oeuvre et par quels acteurs ? Cette subdivision permet de mieux situer les enjeux et limites de la reconnaissance faciale, ainsi que les perspectives de développement et de recherche. Sur le plan des données elles-mêmes, l’enjeu principal concerne leur qualité originale puis sa dégradation potentielle aux stades de collecte et de sauvegarde des images, qui influencent leur utilisation ultérieure. Pour ce qui est des méthodes, les enjeux se cristallisent autour du manque de standardisation et de transparence, aussi bien lors de tâches exécutées par l’être humain que par un système automatique. Concernant les objectifs des tâches de reconnaissance faciale dans les domaines civil et judiciaire, les enjeux gravitent autour de la protection de la sphère privée et des libertés individuelles. Enfin, les principaux défis soulevés par son utilisation comme moyen de preuve au tribunal concernent la communication, ainsi que la standardisation et la validation méthodologique.

The role of forensic science in criminal intelligence becomes more prominent as technological innovations generate an increasing quantity of data to be detected, collected, and processed. In a societal context, where whistle-blowers prompt law-enforcement agencies to be more proactive and raise ethical issues about civil liberties, the concrete benefit provided by this global movement and its integration in policing need to be better understood. Thus, this article focuses on the use of forensic case data in criminal intelligence, especially for the detection of crime patterns. By taking into account a long-term collaboration between School of Criminal Justice and a regional crime analysis unit in Switzerland, the development of a forensic intelligence approach is studied. This approach seeks to deliver intelligence, with the means of crime data stemming from traces and explained/supported by criminological theories. This integration of forensic data in criminal intelligence is perhaps a step toward forensic whistleblowing'.

This article presents the methodological process and the main findings of a research on crime displacement between two cantons (states) of Switzerland from 2009 to 2012. Two analytical axes have been considered: displacement of crime incidents on the one hand, and displacement of offenders for offences against the Swiss Criminal Code on the other hand. Data were provided by police statistics of the two cantons involved and by a regional crime intelligence database, supplemented by documentary analysis, interviews and field observations. Measures of crime displacement were realized with variation in crime rates, difference-in-differences estimation and weighted displacement quotient applied to a selection of offences. Findings suggest the presence of a regional crime displacement of burglary, pickpocketing, simple theft and break-in theft in vehicle between 2011 and 2012. Displacement of offenders seems to follow this trend as well and highlights a small proportion of inter-regional offenders, but who are extremely prolific. These results and the study of displacement in general shed light on the effect of crime reduction strategies and provide many prospective pathways, such as the strengthening of inter-regional collaborations between the police forces in a federal State, and the systematic use of crime intelligence to reduce mobile criminal activities.

The extension of traditional data mining methods to time series has been effectively applied to a wide range of domains such as finance, econometrics, biology, security, and medicine. Many existing mining methods deal with the task of change points detection, but very few provide a flexible approach. Querying specific change points with linguistic variables is particularly useful in crime analysis, where intuitive, understandable, and appropriate detection of changes can significantly improve the allocation of resources for timely and concise operations. In this paper, we propose an on-line method for detecting and querying change points in crime-related time series with the use of a meaningful representation and a fuzzy inference system. Change points detection is based on a shape space representation, and linguistic terms describing geometric properties of the change points are used to express queries, offering the advantage of intuitiveness and flexibility. An empirical evaluation is first conducted on a crime data set to confirm the validity of the proposed method and then on a financial data set to test its general applicability. A comparison to a similar change-point detection algorithm and a sensitivity analysis are also conducted. Results show that the method is able to accurately detect change points at very low computational costs. More broadly, the detection of specific change points within time series of virtually any domain is made more intuitive and more understandable, even for experts not related to data mining.

Despite the predominant role played by Internet in the distribution of doping substances, little is currently known about the online offer of doping products. Therefore, the study focuses on the detection of doping substances and suppliers discussed in Internet forums. It aims at having a comprehensive understanding of products and sellers to lead an operational monitoring of the online doping market. Thirteen community forums on the Internet were investigated and one million topics were extracted with source code scrappers. Then, a semantic analysis was conducted with a semi-automatic process to classify the relevant words according to doping matters. Additionally, the ranking of doping products, active substances and suppliers in regards to the number of contributors to the forums were established and analyzed over time. Finally, promotion methods of suppliers were evaluated. The results show that anabolic androgenic steroids, used to enhance body image and performance, are the most discussed type of products. A temporal analysis illustrates the stability of the most popular products as well as the emergence of new products such as peptides (e.g. CJC-1295). 327 suppliers were detected, mostly with dedicated websites or direct sales by e-mail as selling methods. Globally, the implemented methodology shows its ability to detect products and suppliers as well as to follow their temporal trends. The intelligence will serve the definition of online monitoring strategies (e.g. the selection of appropriate keywords). Additionally, it also allows the adjustment of customs inspection strategies and anti-doping analysis by monitoring the popular and emerging substances.

Grouping events having similarities has always been interesting for analysts. Actually, when a label is put on top of a set of events to denote they share common properties, the automation and the capability to conduct reasoning with this set drastically increase. This is particularly true when considering criminal events for crime analysts, conjunction, interpretation and explanation can be key success factors to apprehend criminals. In this paper, we present the CriLiM methodology for investigating both serious and high-volume crime. Our artifact consists in implementing a tailored computerized crime linkage system, based on a fuzzy MCDM approach in order to combine spatio-temporal, behavioral, and forensic information. As a proof of concept, series in burglaries are examined from real data and compared to expert results.

Grouping crimes having similarities has always been interesting for analysts. Actually, when a set of crimes share common properties, the capability to conduct reasoning and the automation with this set drastically increase. Conjunction, interpretation and explanation based on similarities can be key success factors to apprehend criminals. In this paper, we present a computerized method for high-volume crime linkage, based on a fuzzy MCDM approach in order to combine situational, behavioral, and forensic information. Experiments are conducted with series in burglaries from real data and compared to expert results.