An improved adaptive background modeling algorithm based on Gaussian Mixture Model

Background subtraction is a typical method for real-time segmentation of moving regions from a video sequence. Numerous approaches have been proposed to this problem, which differ in the type of background model. The Gaussian mixture model (GMM) is one of the best models to model a background scene with repetitive motions. However, the large amount of computation has limited its application. In addition, it has difficulty in segmenting slow moving objects and objects that stop for a while during moving. Based on the Gaussian mixture model, this paper presents an improved adaptive background modeling algorithm. A model number adaptive method is used in the algorithm to decrease the amount of computation and an updating method with adaptive learning rate is proposed to accurately segment the objects that move slow or stop for a while during moving. The results which demonstrate the performance of the algorithm are also shown in this paper.