摘要

Video quality is an intuitional and efficient guideline for quality of service of the whole communication system and quality of experience from the end users' viewpoint. In this paper, we present a no reference video quality assessment metric using inter-frame redundancy of video sequences. Human eyes are adapted to extract useful information from viewed video, so we determine these regions of interests through temporal motion information and spatial complexity information. The quality of the frame is compared to its neighboring frames due to the nature of live streams. Finally, the whole video quality is based on these frames' quality. Our metric is focus on decoded video from the end-users' perception and attempt to be used for in-service monitoring and controlling. The experimental results obtained from VQEG Phase-I database indicate its efficiency and potential. Additionally, we conclude two types of factors that influence video quality, and two types of roles of video quality assessment modules during video communication systems.

全文