Research Of Quality Evaluation Technology Of User Experience For Mobile Video And Virtual Reality

Due to a wider bandwidth of mobile communication network,a proliferation of video telephony service has occurred.Apparently operators are highly concerned about the quality of experience of the video telephony perceived by end users in order to conduct the network resource management and optimization.Meanwhile the 5th Generation mobile communication technologies are becoming mature,newly developing services spring up in a large number.Virtual reality service has achieved explosive growth in recent years.As the first kind of booming service in virtual realtiy,panoramic video already has considerable number of users.But research on the quality of experience in this field is still in initial stage.The main contributions of this paper are listed as follows:Firstly,for Vo LTE video telephony service,we set up a Vo LTE video telephony test system in our laboratory.Using this,we capture network parameters in real network.Then we reconstruct original videos and reduced videos basing on captured packets and then set up a Vo LTE video telephony database.Subjective test has been conducted on this database.At last,empirical results are used to derive analytical quality assessment models as a function of time,we introduce the effect of the viewer response to time-varying video quality for video telephony into the existing models based on ITU-R G.1070.Through applying exponentially-weighted Minkowski summation,we could find that the correlation between the prediction of the improved model and subjective scores becomes much better.The Pearson correlation coefficient could reach up to 0.96.On the other hand,for panoramic video service,we set up a virtual reality panoramic video live system.Based on this system,we simulate the panoramic video live service by adding network damage such as packet loss and delay with different bitrates in our local network.Then we get reduced videos in different network conditions,together with source videos,which make up the panoramic video live database.We divide these videos into 3 scenes,sports,convention and concert and set up 4 subjective indicators,video quality,immersion,dizziness,quality of experience in our subjective test.Moreover we make further research on the subjective results,exploring the impact of bitrate,packet loss rate and delay on quality of experience and the coefficient between different subjective indicators.Finally,we propose an objective quality assessment model based on neural network and validate its performance using video sequences in three scenes.The Pearson correlation coefficients in three scenes are all over 0.9 which means an excellent performance.