Segmentation becomes a difficult task when the background illumination changes.In this paper

we apply a Bayesian learning method into video segmentation.The constantly changing background has been modeled at the pixel level.The feature vector for each pixel is represented with a discrete probability distribution function.The histogram colors and co-occurrence vectors have been calculated.Bayesian learning has been used to obtain these probability distribution functions from the video image inputs.The experimental results indicate that the proposed approach is able to learn a complex background of which the illumination changes either gradually or suddenly.