Approved projects in the field of "Pattern recognition"

The joint projects presented in this section are devoted to the development of innovative techniques for automatic capturing, targeted recognition and processing of data from a variety of sources, such as cameras and sensors. Improving the evaluation of such data can provide security services with a sound basis upon which to assess potential risks earlier and more accurately. The projects focus on preventive solutions, giving particular consideration to data protection requirements and social acceptance of the technologies concerned.


Boyan Dimitrov /

Completed joint projects:

ADIS: Visual pattern classification for automated detection of situations requiring intervention

Animation eines ausgewiesenen Bereiches, in dem der Fahrgast, nach individuellem Bedarf, besonderen Schutz in Anspruch nehmen kann.

Fraunhofer IPK

Funding codes 13N10977 to 3N10979

Media reports about violent attacks in underground stations are a common occurrence and reflect, among other things, a greater need for safety measures in these public places. In many places, video cameras have already been installed at underground station entrances and in the underground trains themselves. However, the sheer mass of video material often means that the images cannot be evaluated until a later stage. With these aspects in mind, the ADIS project developed a solution for designated areas of stations, which are fitted with cameras with pattern recognition systems. This enables aggressive behaviour or medical emergencies to be recognised and thus immediate action to be taken in hazardous situations. Such action could take the form of de-escalating announcements or deployment of security staff. The system makes it quite clear to members of the public that it is their individual choice whether they move to a designated area for their own safety or not.

More information  (only available in German)


APFel: Analysis of subjects’ movements at airports using forward and backward tracking in video data streams

Flughafenterminals sind in Anbetracht anhaltender Terrorismusgefahren Bereiche mit einem hohen Sicherheitsbedarf.

TU Ilmenau / Hannover Airport

Funding codes 13N10795 to 13N10801

Security requirements are high in airport terminals, which is why more and more video surveillance cameras have been installed in recent years. The APFeL project carried out research on a system to support video data evaluation by enabling control room staff to flag suspicious individuals so they can be tracked more easily, using multiple cameras, as they move around the airport. By comparing movements with typical movement patterns, for instance, the new system makes it possible to predict those individuals’ subsequent movements, thus making it much easier to assess their risk potential at an early stage. Due to the technical features of the solution developed, especially with regard to the handling of personal data, the project also concentrated on aspects related to data protection and basic rights.

More information  (only available in German)


ASEV: Automatic situation interpretation for event-triggered video surveillance

Flughafenterminals sind in Anbetracht anhaltender Terrorismusgefahren Bereiche mit einem hohen Sicherheitsbedarf.

TU Ilmenau / Hannover Airport

Funding codes 13N10716, 13N10718, 13N10719 and 13N11040

Criminal or terrorist activities and interference in the loading/unloading or maintenance of aircraft can jeopardise airport security to a huge extent. The ASEV project therefore worked on a model for a semi-automated system with which airport staff can better assess potentially dangerous situations by means of event-triggered video evaluation. The objective was to reduce the false alarm rate and the need to constantly observe people on the airfield “manually”.

More information  (only available in German)


CamInSens: Distributed, networked camera systems for in situ detection of risk situations caused by individuals

Beispielhafte Überwachung mit dem CamInSens – System

Leibniz Universität Hannover

Funding codes 13N10809 to 13N10814

The CamInSens project focused on the development of a practical, legally compliant and intelligent video system that alerts the operator to potential risk situations immediately and automatically. The developed solution for evaluating image sequences enables suspicious movement patterns to be detected and the camera to be controlled so as to keep track of relevant situations. Right from the beginning of the joint project, the team examined the specific legal issues involved in order to ensure the solution was legally compliant. This examination ranged from a detailed analysis of the legal requirements to a legal evaluation of the demonstrator.

More information  (only available in German)


DigiDak: Digital fingerprints

DigiDak befasst sich mit der besseren Erfassung von Fingerspuren.

Funding codes 13N10816 to 13N10822

Although fingerprint analysis has greatly improved investigators’ success rates in recent years thanks to the use of pattern recognition systems, fingerprint collection is currently an exclusively non-automated process. This project explored new approaches for making fingerprint detection easier and making collection both more exhaustive and speedier – in luggage and freight security checks, for instance. The findings were combined in a software demonstrator, which is intended to facilitate time-efficient and automatic collection of fingerprints, even at large crime scenes, while ensuring compliance with constitutional and data protection law.

More information  (only available in German)


INBEKI: Interaction-triggered image data analysis to combat child pornography

Bei der Bekämpfung von Kinderpornografie sind die Ermittlungsbehörden oft mit einer Flut von Daten konfrontiert.

Landeskriminalamt NRW

Funding codes 13N10783 to 13N10787

Due in particular to the web and high-volume data storage devices, authorities investigating registered cases of child pornography (which are on the rise) are faced with overwhelming amounts of data. This makes systematic manual evaluation extremely difficult and slow. The INBEKI project aimed to create an integrated system solution to provide automated support for the process of evaluating confiscated data. To this end, it used new image and video analysis methods to make it easier to find and recognise offenders, victims, objects and crime scenes. To ensure the system as a whole was fit for purpose, the demonstrator was evaluated in field tests using real data.

More information  (only available in German)




MuViT: Pattern recognition and video tracking: socio-psychological, sociological, ethical und legal analyses

Bei der technischen Entwicklung von Mustererkennungssystemen ist es wichtig, rechtliche und ethische Fragen mit zu lösen.

Eberhard Karls Universität Tübingen

Funding codes 13N10958 to 13N10962

Automated pattern recognition technologies such as video tracking offer great potential for providing security personnel with effective support when analysing video surveillance material. Such “intelligent” video systems do, however, raise certain social, ethical and legal issues, which also need to be addressed when developing new technology. The partners on the MuViT joint project therefore observed several projects focusing on pattern recognition to develop a list of criteria as a guide for determining the conditions under which pattern recognition systems can comply with social, ethical and legal requirements.

More information  (only available in German)


SICURA: Security screening using x-ray analysis

Röntgenbild eines Handgepäcks

Smiths Heimann GmbH

Funding codes 13N11122 to 13N11125

Airports use x-ray scanners to detect explosives and weapons in passenger luggage. But weapons, in particular, have components that come in different shapes and sizes, making it difficult to detect them automatically every time. Consequently, suspicious hand luggage requires visual evaluation of the images and a manual inspection by security personnel. However, the fewer additional manual checks that are necessary, the faster passengers can be processed. With this in mind, the SICURA project developed a software solution for detecting dangerous items in luggage automatically. The software enables the security screening personnel to automatically identify various objects in luggage items, thereby supporting them and lessening their workload.

More information  (only available in German)


VASA: Visual Analytics for Security Applications

Funding codes 13N11252 to 13N11256, 13N11258 and 13N11259

Decision makers in the areas of disaster preparedness and crisis management have to process a large amount of information in a very short amount of time. When considering that information, they also need to take into account the interdependency of important infrastructures in security-critical situations. The German-US VASA joint project was therefore seeking to improve disaster preparedness and crisis management by employing visual analytics. The aim was to produce a demonstrator that will enable persons such as control centre or operation managers to grasp complex crisis situations quickly with the aid of visual evaluations. This allows action to be taken quickly in order to prevent breakdowns or to minimise their impact.

More information (only available in German)