Testing video analytic algorithms
Posted on : 19-01-2011 | By : Yuri Vashchenko | In : Uncategorized
0
A modern video analytic system depending on business/customer requirements should work in different situations/conditions. Complex, noisy background with many different objects/textures, changing lighting conditions, shadows, lack of light, weather conditions (for outdoor system installations) like rain, snow, fog and others, motion blur, camera movements, cameral sensor quality, camera resolution, camera focus issues, camera internal optimizations, color temperature, end many other factors make development of the good object recognition software a challenging, almost impossible task. In addition, the usual requirement is that the system should work in real time, which makes this task even more difficult.