Two participants of VANAHEIM interviewed by French magazine of science popularization S&V.
Science & Vie n°1111, Avril 2010.
Starting from the same findings that led to VANAHEIM building (i.e. that 1. most of the time videos from large surveillance network are never watched and 2. the reality regarding the effective reliability of audio/video analytic is that it is still very much in its infancy), the article expresses objective criticisms against videosurveillance, and points out the need for more research in the direction of behavior recognition, as proposed in VANAHEIM.
VANAHEIM Project, funded by the European Commission, should provide in four years to Paris and Turin subways' responsibles new means of automatic analysis of images captured by their videosurveillance networks.
Jean-Marc Odobez, specialized in computer vision at IDIAP Research Institute and involved in this project, acknowledges the difficulty of the task: "It quickly conducts to very high levels of abstraction in algorithms, to successfully identify and monitor various objects and know what to infer from a scene", he said.
Get the videosurveillance article from Science & Vie n°1111, Avril 2010 (in french).