A Video Quality Assessment Model Based on the MPEG-7 Descriptor

Our research is focused on examining the video quality assessment model based on the MPEG-7 descriptor. This model consists of two parts: "Frame Quality estimation processing" and "Video Quality estimation processing". The estimation of Video Quality in the proposed model uses five values (average value, worst value, best value, standard deviation and frame rate) from the estimation Frame Quality and the input video sequence. Two coding methods (WMV9 and H.264) are used to verify the proposed model's presumption accuracy. As a result, Video Quality estimation has a high presumption accuracy (correlation : 0.94, average error : 0.20, maximum error : 0.68 and outlier ratio : 0.23).