ABSTRACT
In Chapter 1, we provided a survey of the traditional background models for videosurveillance and how to evaluate them using measures evaluation and datasets. However, no traditional algorithm today seems to be able to simultaneously address all the key challenges that we meet in videos. This is due to two main reasons:
1. Lack of scientific progress, that is, there are mainly improvements of the state-ofthe-art methods such as Mixture of Gaussians, and other representation models are sometimes investigated insufficiently.
