Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Face- and appearance-based person identification for forensic analysis of surveillance videos

 
: Herrmann, C.; Metzler, Jürgen; Willersinn, Dieter; Beyerer, Jürgen

:
Volltext urn:nbn:de:0011-n-3625774 (698 KByte PDF)
MD5 Fingerprint: 8c8810da703a0983d07f7fe4f7bb3884
Erstellt am: 15.10.2015


Beyerer, Jürgen (Ed.); Meissner, Andreas (Ed.); Geisler, Jürgen (Ed.) ; Fraunhofer-Institut für Optronik, Systemtechnik und Bildauswertung -IOSB-, Karlsruhe:
10th Future Security 2015. Security Research Conference. Proceedings : September 15 – 17, 2015, Berlin
Stuttgart: Fraunhofer Verlag, 2015
ISBN: 978-3-8396-0908-8
ISBN: 3-8396-0908-9
S.441-444
Security Research Conference "Future Security" <10, 2015, Berlin>
Englisch
Konferenzbeitrag, Elektronische Publikation
Fraunhofer IOSB ()

Abstract
The increasing availability of surveillance cameras is both an opportunity and a challenge for forensic crime investigation. For serious crimes, video footage offers a widely accepted opportunity to identify criminals and reconstruct the events to find further offenders. Because this is often a highly manual task, automated video analysis methods are welcome to efficiently handle the growing amounts of video data. The research project MisPel (Multi-Biometriebasierte Forensische Personensuche in Lichtbild- und Videomassendaten) funded by the German Ministry of Education and Research addressed this field by creating and combining several automated video analysis tools into a demonstration system. A forensic analysis system is always controlled by a human operator who selects which data are to be analyzed and what kind of analysis should be performed. Here, the focus will be on one specific part: the identification of persons to find all occurrences of an offender in the video data. Assuming that an offender has caught the operator’s attention, the aim is to assist the operator by finding further occurrences of this particular person in the relevant video data. This contribution focuses on the technical part of the data extraction.

: http://publica.fraunhofer.de/dokumente/N-362577.html