Removal shadow with background subtraction model ViBe algorithm

Removal shadow is an important step in many video surveillance applications. But shadow pixels belong to motion foreground and often misclassified as object during segmentation and tracking. Currently are many algorithms that deal with shadow in different color spaces in papers. However, it is not effective to determine the shadow segmentation and remove the shadow. In this paper, we study the invariant color features with the background model of Visual Background Extractor (ViBe) elimination moving object shadow, and repair pixels must be foreground not be shadow part, and this pixels enhance foreground mask convexity.