Objective video quality assessment for tracking moving objects from video sequences
Engin Mendi, Mariofanna G. Milanova, Yufeng Zhou, John R. Talburt · International Conference on Signal Processing · 2010
Video quality assessment has a great importance in several image processing applications. Recently, various objective video quality metrics have been proposed in order to predict the human visual perception and to achieve high correlation with the human perception of the image quality. In this paper, a novel objective quality metric is proposed for tracking moving objects in video sequences. The proposed metric particularly considers the moving objects in video sequences as visually important content. Foreground masks are produced by background subtraction based an approximate median filter. Existing metrics are then modified by the weighting factors of the foreground masks. Our results show that our metrics have better performance than existing objective metrics.