Video quality estimation for packet loss based on No-Reference method

Objective video-quality assessment can be divided into Full Reference (FR), Reduced Reference (RR), and No Reference (NR) methods. Among them NR quality assessment method does not require original video sequence, so it is suitable for the real-time video streaming service. This paper presents an NR-based estimation of mean square error (MSE) for video quality degradation due to various pattern packet losses in packet video network environment. Our proposed algorithm predicts initial error due to packet loss in each frame by using motion information of received data. Proposed method shows high Pearson correlation (more than 95%) to match actual MSE.