Attention Score: Objective Measure of Attentiveness in Immersive Omnidirectional Videos
Jay Bhanushali, Achsah Steffi John, Manivannan Muniyadi · 2022
The application of Virtual Reality (VR) in education and entertainment has shown to increase the overall learning experience of the users and inspire them in unique ways. Identifying and analyzing the viewing and attention patterns of users is extremely important for improving their engagement with the content in VR. In this paper, we introduce two new objective metrics: Attention Score (ATNS) and Maximum Attention Span Score (MASS) to quantify the attention of a user to a Point of Interest (PoI) in a VR scenario. The metrics are computationally inexpensive and require no additional hardware. As part of our work, we present ReptilesVR, a VR-based edutainment and cinematic zoo experience developed using omnidirectional stereo videos to provide immersive close encounters with wildlife. We conducted a study with 25 participants to quantitatively measure the overall experience of users in ReptilesVR using the two newly defined metrics: ATNS and MASS. Also, the emotional engagement of the users was measured qualitatively using a questionnaire on a Likert scale of 1–10. We observed that scenarios with activity in close proximity to reptiles have positive correlation between the proposed metrics with the subjective scores of presence (μ = 8.58, σ = 1.28) and immersion (μ = 9.0, σ = 1.14) as reported by participants in the questionnaire. This work shows that the metrics can be useful to find the features that draw a user’s attention to certain points of interest (PoIs) in a scene. This will aid in the enhancement of the overall user experience and the development of more compelling immersive VR applications.