Real-time Video Quality Assessment System Based On Human Eye Masking Characteristics

In recent years, with the rapid spread of the Internet, online video service has almost penetratedinto various areas around the world, it has brought great convenience to people’s lives, but videocompressed encoding and network transmission can cause varying degrees of distortion, soestablishing a set of accurate, real-time assessment method of the video quality of the modelbecomes very useful.This paper studies and analyzes the results of the relevant research at home and abroad aboutthe video quality assessment, elaborate on the related technologies, and analyze their advantagesand disadvantages further. Then, on this basis, the network video quality assessment model:HVQAM is presented. It is according to human visual characteristics for intra-regional weighted,drawing frame image quality assessment by support vector machines (SVM), and then according tothe human visual system overall frame image the overall weighted average is conducted, in order tomake the assessment results of the model is more consistent with the perception of the human eye.Finally, prototype system is designed and developed based on HVQAM, and then the results isdisplayed and analyzed. This system can be used for automated real-time assessment of the video,and the validity as well as the advanced nature of this model is verified from three aspects of theevaluation results and objective quality consistency, CPU utilization and memory usage rate.The system has been audited by Huawei Company, and has already been used for mediaanalysis unit (MAU) in network video surveillance products.