Quality Assessment of Virtual Reality Videos

360-degree spherical images/videos, also called virtual reality (VR) images/videos, can provide an immersive experience of real scenes in some specific systems. This makes it widely used in VR games, sporting events and VR movies. However, due to its high resolution, it is so difficult to transmit, compress or store VR images/videos. Therefore, it is significant to study how noise affects the quality of VR images. To this end, this paper builds a VR video database, and carries out subjective and objective experiments on them. Specifically, first, six standard panoramic videos are processed by inputting three kinds of distorted types to establish a VR video database which comprises 96 videos. Second, we utilize the Double Stimulus Injury Scale (DSIS) for subjective experiments. All subjective scores are from 20 non-professional viewers. Third, we utilize 6 existing objective metrics to validate our database. Finally, experimental results demonstrate that the established VR database is suitable for subjective and objective quality evaluation of VR video. Our work has alleviated the problem of missing VR databases.