Network Video Service Quality Assessment Research

With the increase of network bandwidth in recent years, real-time video services such as streaming media business, video session business and video teleconferencing services developed rapidly. The accurate assessment to these video services plays an important role in the business operation as it not only informs users with the service level of video business but also provides a scale to improve service level to services provider. Online monitoring of the services, especially in the actual operation of the business, could dynamically obtain the quality state of business operations. It also has a strong practical value for it could detect the quality problem during the services operation in time. This research is mainly about how to achieve more effective video quality evaluation on the network side.Firstly, this paper briefly introduces the traditional video quality evaluation algorithm and then focuses on several classic network side video no-reference evaluation algorithms. This paper innovatively proposed a network side video services no-reference quality evaluation model used in3G wireless Internet network. The model can achieve more accurate quality evaluation scores by comprehensive analysis of three aspects of video business:source codec damage, packet loss damage, delay jitter damage. This model also introduces a concept of sliding time window which is used to reduce the volatility of the data, so that to be closer to subjective feelings. The accuracy of the models is tested with video data and verified through the analysis of correlation parameters and graph comparing.Then, this paper establishes several associated modules. By connecting these modules can achieve the functions such as capturing, assessing and storing for the video stream which increases the portability of the video assessment algorithm. Each module can completed real-time video quality evaluation function in parallel through multi-threaded program architected and effective data structures designed. This article uses the video test sample for each module of the functional tests to verify its usefulness and effectiveness.Finally, this paper implemented a network side video quality evaluation system. Establishment of the system based on the video quality evaluation module while combined a statistical module and a graphical display module, constitute a complete video quality evaluation system which can be applied at the network side. The display system can display different scores by meanings and the data can be updated in real-time. This article tests the operation of the system and achieves satisfactory results.