KAIST IVY LAB 360-degree Video Database with Different Motion Patterns

VR Video Test Datasets

The videos provided below are test datasets from Youtube to evaluate the performance of the proposed method in the paper and measure the VR sickness in subjective assessment.  They were captured during driving on a road and had different motion patterns for three different scenarios. Average motion in the video 1 was slow and the video 2 had moderate motion velocity.  In the video 3 captured in racing car, its average motion was very fast.

VR Video 1 (Slow motion) : [LINK]

VR Video 2 (Moderate motion) : [LINK]

VR Video 3 (Fast motion) : [LINK]

Corresponding Subjective VR Sickness Score Datasets

The number of subjects participated in subjective assessment experiment is 15. In experiments, 16-item SSQ was used to measure the degree of VR sickness in watching the VR video contents. For the assessment of VR sickness, we asked participated subjects to complete the SSQ sheet before and after being exposed to three VR video contents. The subjects used a discrete scale divided into four levels in order to grade the VR sickness. The labels of SSQ were 'None', 'Slight', 'Moderate', and 'Severe'. For each symptom, the score were 0 for 'None', 1 for 'Slight', 2 for 'Moderate', and 3 for 'Severe'. Finally, a total SSQ score was calculated by combining every partial scores for each symptom with the weight, which was set to 3.74.

Subjective VR sickness score datasets : [Link]

Reference

If you use the database, please cite as : 

[1] Kim, H. G., Baddar, W. J., Lim, H. T., Jeong, H., & Ro, Y. M. (2017, November). Measurement of exceptional motion in VR video contents for VR sickness assessment using deep convolutional autoencoder. In Proceedings of the 23rd ACM Symposium on Virtual Reality Software and Technology (p. 36). ACM.