Pixel-based colour contrast for abandoned and stolen object discrimination in video surveillance

A novel approach is proposed for discriminating between abandoned or stolen previously detected stationary foreground regions in video surveillance. It is based on measuring the colour contrast of the contour of the stationary object under analysis at pixel level. Two contrasts are computed by analysing such a contour in the current and background frames. Then, both are combined for performing the discrimination. The experimental results over a heterogeneous dataset containing real scenarios demonstrate that this approach outperforms the related literature and greatly reduces the computational cost of the discrimination task, allowing real-time operation.