A Fast Convergent Gaussian Mixture Model in Moving Object Detection with Shadow Elimination

Gaussian mixture model is a commonly used background modeling method in moving object detection. Gaussian mixture model has a strong adaptivity to various complicated backgrounds, but converges slowly and lacks shadow detection capability. In this paper, we propose an improved Gaussian mixture model which models background and foreground at the same time, accelerates convergence when moving objects suddenly stop and completes object detection with shadow detection simultaneously. Experimental results show that the proposed improved Gaussian mixture model achieves better results in shadow detection and converges more quickly when a sudden stop happens.