ACR360: A Dataset on Subjective 360° Video Quality Assessment Using ACR Methods
Majed Elwardy, Hans‐Jürgen Zepernick, Yan Qing Hu, Thi My Chinh Chu · 2023
The recent advances in immersive technologies have been essential in the development of a wide range of novel standalone and networked immersive media applications. The concepts of virtual reality, augmented reality, and mixed reality relate to different compositions of real and computer-generated virtual objects. In this context, 360° video streaming has become increasingly popular offering improved immersive experiences when viewed on a head-mounted display (HMD). An important component in the development of novel immersive media systems are subjective tests in which participants assess the quality of experience of representative test stimuli. In this paper, the annotated ACR360 dataset is presented which is publicly available on GitHub. The ACR360 dataset contains a wide range of psychophysical and psychophysiological data that was collected in Subjective tests on 360° video quality. The test stimuli were shown on an HMD and rated according to the absolute category rating (ACR) and modified ACR (MACR) methods. To support an easy exploration and utilization of the ACR360 dataset by the research community, its structure on GitHub is described and a comprehensive illustration of analysis options are provided for each data category. The ACR360 dataset may be used for conducting meta-analysis in combination with other datasets to improve precision and to pursue research questions that cannot be answered by an individual study.