Re-identification : On the perception of human motion for surveillance and forensic applications

Because of the unprecedented number of crimes and terror attacks as well as the vital need to provide safer environment, a surge of concerns has emerged in many countries to ensure the safety of their citizens via the use of advanced surveillance technology. It is no wonder that many cities have deployed a number of biometric technologies within their surveillance systems to ease the management and monitoring from a large number of cameras. Biometrics can be of benefits not only for identify recognition, but it can be play a vital role for tracking people people over a distributed camera network with disjoint views. Gait biometrics is considered the only suitable solution in surveillance scenarios to re-identify people between different cameras due to the fact that the gait motion can be captured from a distance even with low-resolution besides the non-invasive property of gait biometrics.