An Improved ViBe Algorithm Based on Visual Saliency
Peng Li, Yanjiang Wang · 2017 International Conference on Computer Technology, Electronics and Communication (ICCTEC) · 2017
In order to solve the problems of "ghost" effect and noise interference of the classical Vibe algorithm in moving object detection, an improved Vibe algorithm based on visual attention mechanism was put forward. In this algorithm, the two-dimensional entropy and saliency of any frame are firstly calculated, by which the adaptive background updating factor is derived. Then, the background model can be adaptively updated according to the changes of background between adjacent frames. And the visual saliency is also used to suppress and eliminate the ghost effect rapidly. Comparison of experimental results show that in the absence of prior knowledge of the moving targets, the improved algorithm can eliminate the ghost more quickly, and the foreground targets can also be detected more accurately.