| With the development of communication technology and the popularization of smart terminal devices,for instance,mobile phones and tablets,people's demand for video is increasing day by day,and the wireless video service has become one of the most popular multimedia applications.However,the dynamic change characteristics of the wireless channel can easily cause video data loss,resulting in degradation of video quality and affecting the Quality of Experience(QoE).Therefore,it is necessary to accurately assess the wireless video quality in real time,and provide guides for the adjusting the parameters of the encoder/decoder online,thereby improving the QoE.The factors influencing the QoE are firstly analyzed in this paper,including system factors,context factors and human factors.Then,according to the different classification methods of objective video quality assessment,the up-to-date objective video quality assessment methods are analyzed and compared.Aiming at the problem of video quality degradation caused by the events including compression coding,rate change and rebuffering etc.,combined with human visual characteristics,two objective wireless video quality assessment methods based on human visual characteristics are proposed in this paper,and the applications of objective wireless video quality assessment methods are studied and the applications of wireless video objective quality assessment methods are studied.A no-reference mobile video quality assessment based on natural statistics is proposed.Considering the influence of compression and wireless packet-loss on mobile video quality in wireless networks,analyzing the space-time perceptual statistics of the differences between video adjacent frames,a no-reference mobile video quality assessment called NMVQA based on natural statistics is proposed.Firstly,the influence of various video distortion types on the statistical characteristics of difference coefficients between video adjacent frames are analyzed in terms of the natural statistical regularities of video frame difference.Secondly,the temporal change of the distribution parameters with respect to the products of adjacent frame differences computed along horizontal,vertical and diagonal spatial orientations are calculated.Finally,the distortion degree of the mobile video is measured by the correlation between the multi-scale temporal changes of statistical characteristics of difference coefficients between video adjacent frames.Experimental results in the Laboratory for Image & Video Engineering(LIVE)mobile video database show that the results of NMVQA are well consistent with subjective assessment results,and can reflect human subjective feeling well.NMVQA can evaluate the performance of real-time online adjustment of the source rate and wireless channel parameters.A spatial and temporal feature-based reduced reference quality assessment for rate-varying videos in wireless networks is proposed.For the influence of the dynamic rate change of the wireless network on the terminal video quality,based on the extraction of the temporal and spatial domain features of quality distortion degree of rate-varying video in wireless network,combined with the recency effect,a spatial and temporal feature-based reduced reference quality assessment for rate-varying videos in wireless networks called STRQAW is proposed.Firstly,utilizing the orientation selectivity mechanism of the human visual system(HVS),the histogram of the orientation selectivity-based visual pattern in each frame is extracted as the spatial feature.The histogram similarity between the rate-varying video and the original video is computed as the spatial metric.Secondly,the temporal variation of the DCT coefficients of the consecutive frame differences is extracted as the temporal feature.The temporal variation similarity between the rate-varying video and the original video is calculated as the temporal metric.Finally,we take into account the recency effect and assess the overall quality by combining the temporal and spatial metric.The experimental results tested in the LIVE mobile video quality assessment database show that STRQAW is well consistent with the subjective assessment results,which means it reflects human subjective feelings well and it provides an evaluation for adjusting rates in real time.STRQAW can be used to provide guides for content providers and network operators to adjust rates and channel parameters to improve end-users' satisfaction.An SSIM based rate clustering recognition algorithm for wireless video is proposed.In wireless video transmission process,due to the complexity and variability of wireless communication channels,the rate is adjusted to match the dynamic wireless channel,which will affect the video quality and QoE,it is necessary to dynamically recognize the rate variation.According to the distribution characteristics of video frame quality,an SSIM based rate clustering recognition algorithm for wireless video called SBRCA is proposed to recognize different rates for wireless video transmission.SBRCA firstly uses SSIM between each pair of original and terminal decoding frames as the feature to quantify video frame quality,and it computes the similarity measure between two adjacent consecutive SSIM values.Then based on the principle of the binary tree,the frame interval is recursively split by the frames with smallest value of similarity measure in each splitting process.Finally,the different video bitrates can be clustered by SBRCA,and the bitrate variation can be recognized.The results show that the proposed algorithm can recognize the change of video bitrate by analyzing video frame quality and it is well consistent with the real bitrate change in wireless video transmission.A continuous prediction model of wireless video QoE is proposed.Due to the rapid development of communication technology,the demand for mobile video streaming services has exploded,but the limited bandwidth of wireless networks has frequently caused video quality distortion events such as rebuffering event during the video playback,and has decreased the QoE.Predicting the impact of video impairment on QoE can provide guides for improving resource allocation strategies to provide users with higher quality video streaming services.In order to predict the impact of video impairment on the user's QoE,a continuous prediction model of wireless video QoE is proposed.The inputs of the prediction model consist of the video frame quality vectors,the state vectors of the rebuffering event,and the vectors characterizing the memory effect.The output consists of the continuous subjective QoE.The prediction model uses the block-based nonlinear Hammerstein-Wiener model.The experimental results on the LIVE-Netflix mobile video database show that the proposed model has good consistency with the ground QoE and can accurately reflect the human subjective perception on wireless video.The paper contains 54 figures,44 tables and 190 references. |
PDF Full Text Request