Theories And Methods For Wireless Video Quality Assessment

It is pivotal for the new generation of wireless network to evaluate its video quality. Video quality assessment can be widely used in the designing, comparing and application of the communication systems. So it attracts great interest of researchers. However, it is still an unsatisfied and challenging subject for people to evalute the quality of a digital video. Due to its characteristic, traditional methods which can be used are not suitable for wireless application. This paper based on the characteristic of the wireless application and aimed to find the suitable method to evaluate the quality of video.Firstly, the paper introduced the physiological and psychological mechanism of the human visual system (HVS) which played an important role in video and image processing technology. Secondly, the paper systemically analyzed the current theories and methods for the video quality assessment and classified them into off-line-mode and on-line-mode based on their wireless application. By doing so, it would be more visible to find out the advantage or disadvantage of the current theories and methods in their wireless application. Then a general simulation and analysis platform was founded to help the author accumulate all kinds of distorted videos and discover the main factors which contribute to the influence of the video quality. Thus a new model which based on the principle of so call content-pre-detection and activity-division was proposed for the wireless video quality evaluation. By doing activity-division, it would be possible for the system to choose the suitable strategy for the quality evaluation according to the different characteristic of the different video activity regions. While by doing the content-pre-detection, it would be able to take good use of the spatial and temporal information and conbine them to guide the strategy for the video quality assessment. Finally, the paper discussed two types of wireless application of the model which were on-line-mode and off-line-mode, and builded a relationship to map the parameters to the subjective score.Following the research above, the paper used the four metrics which proposed by the VQEG (Video Quality Expert Group) to test and validate the performance of the model. By comparing it to other models, it is proved to have better relationship with the subjective score and have more predominance to be used in the wireless application.