
The re-identification of persons is an important step in camera surveillance systems. The amount of data generated by such systems is often too large to be manually analyzed. Automatic and semi-automatic re-identification approaches can help speed up this process by presenting the operator with likely matches of a query person.
At Fraunhofer IOSB we develop such re-identification methods using deep learning methods. Our approaches are able to quickly search large amounts of data in an automatic or interactive process.