A novel object detection technique for dynamic scene

Object detection in video streams plays an important role in many computer vision applications and background subtraction is the most popular method which compares each new frame to a model of the background. However, the static background is practically impossible and dynamic background makes the perfect object detection very difficult. In the paper, based on the visual background extraction (ViBe) algorithm, we present a new method dealing with this problem. A new update mechanism is present to get more robust model for the dynamic regions. We describe our method in full details and compare it to other background subtraction techniques. Our experimental results show that our proposed method outperformed several state-of-the-art object detection approaches. Moreover, it can process 60 frames per second on an Intel i5 3.1GHz CPU with C++ code.