A dedicated framework for weak biometrics in forensic science for investigation and intelligence purposes: The case of facial information

Following the deployment of strong biometric systems in forensic science (for example, finger/palmprints or DNA), additional weaker biometric data such as facial information, ear or gait, are making their way into police practices and judicial systems. Their introduction is not going without presenting new challenges because of their lower discrimination power (that is, their efficiency at distinguishing individuals). Current biometric systems are designed and deployed as stand-alone applications (operating on their own merit, detached from any other investigative information) and are not fit for purpose when dealing with less discriminating modalities such as faces. We posit in this article that, for these emergent modalities, a different framework, integrated with the policing strategy, is required. The proposed framework is designed to maximize the payoff of these modalities for investigation or intelligence purposes. The number of facial images of non-identified individuals of interest available to police forces is increasing. Their sources go from surveillance cameras, cameras from automated teller machine, personal devices and so on. We analyzed, between 2009 and 2013, data from a regional intelligence platform, used by the crime intelligence units and, show using real case examples, the potential of facial images for crime investigation and crime intelligence.