Estimation of video quality of H.264/AVC video streaming

In this paper the problem of estimation of subjective video quality of distorted video sequences for mobile video streaming is addressed. Widely spreading mobile communication systems and increasing data transmission rates expand variety of multimedia services. One of such services is video streaming. So it is important to asses quality of this service. Consumers of video streaming are humans, and quality assessment must account human perception characteristics. Most of the well-known video quality estimation methods, such as Peak Signal to Noise Ratio (PSNR), Structural Similarity (SSIM), Video Quality Metric (VQM), etc. are based on the presence of full or reduced reference video sequence but not always are good indicators of human perception. In order to estimate user experienced video quality, methods with no reference must be employed. Such existing methods as quality metrics usually use bit-error rate that has low correlation with by human perceived video quality, too. More advanced methods usually require too much processing power that cannot be obtained in handled mobile devices. The proposed no reference video quality estimation method estimates video quality degradation with respect to human perception that is measured as VQM scores. It is also shown that given method is low computation complexity and could be developed basing on H.264/AVC video stream structure. Such method is suitable for implementation in mobile devices with low processing power.