Currently, at low bit rates, the MPEG compression coding can generate some impairments, which can affect the visual video quality. These artifacts such as the blocking, blurring and ringing effects can be exploited in order to design No Reference video quality metric. In this paper, we propose to use an importance map extracted from a region based attention model, to weight distortion measures derived from previous works [1]. This perceptual map is generated from the processed image, combining on the one hand, the simulated results of visual human cell responses and on the other hand, the information of a spatial segmentation. The contribution of these weights is firstly evaluated, in the case of the quality assessment of JPEG and JPEG-2000 compressed images: this perceptual validation allows to assure the relevance of proposed distortion measures. Then, the performance of these combined measures is performed using a database composed of MPEG compressed videos. High correlation between the objective scores of the proposed metric and the subjective assessment ratings has been achieved.
Rémi Barland, Abdelhakim Saadane, "Reference Free Quality Metric using a Region-Based Attention Model for MPEG Compressed Videos" in Proc. IS&T CGIV 2006 3rd European Conf. on Colour in Graphics, Imaging, and Vision, 2006, pp 484 - 488, https://doi.org/10.2352/CGIV.2006.3.1.art00099