Pixel-based statistical analysis of packet loss artifact features

Due to high requirements on network capacity during network transmission, video signals have to be compressed. Compression process and network transmission introduce different compression artifacts and packet loss (PL) artifacts in video signals, respectively. These artifacts degrade visual quality of video signal and thus video quality has to be continuously measured and monitored in order to assure the target quality of service. Regarding PL artifacts, decoder's post-processing algorithm tries to mitigate or completely remove the visual impairments caused by the PL. Consequently, depending on error concealment method, different types of PL artifacts may be formed. In this paper we made a categorization of PL artifacts. The pixel-based statistical analysis of each PL artifact type is performed. The results show that the interesting statistical features can be extracted for the particular PL type, which can be further exploited in PL artifact detection algorithms and video quality evaluation algorithms.