Python 3.6.10
Tenorflow-gpu 1.2.0
Matlab R2020a
In this study, we first develop a useful CG animation subjective video quality database for the validation of corresponding VQA algorithms. It consists of 27 reference videos and 397 distorted videos. The distortion types include five compression-based distortion types and one transmission-based distortion type. All the videos in are High Definition (HD) content and above. More details of the database are shown as follows:
| Number of Reference / Distorted Videos | 27 /397 |
| Sources | animated films, games |
| Resolution | 1270x720 (720p), 1920x1080 (1080p), 3840x2160 (4K UHD), 4096x2160 (DCI 4K) |
| Distortion Types | ● AVC/H.264 compression: qp=22, 32, 42, 50 ● HEVC/H.265 compression: qp=32, 42, 47, 50 ● MPEG-2 compression: q=14, 31 ● intraframe-only compression byMJPEG codec ● wavelet-based compression by Snow codec ● additive white noise: σ=0.003, 0.005, 0.01 |
| Frame Rate | 24fps, 30fps, 60fps |
| Subjective Test Setup | ITU-R Recommendation BT.500 |
| Subjective Test Method | Single Stimulus (SS) |
| Subjective Score Data | mean opinion score (MOS) |
| Display | DELL U2720Q |
| Viewing Distance | 2.5 picture heights |
| Number / Age / Famale Percentage of observers | 25 / 22～35 / 44% |
The corresponding newly established database is available at https://pan.baidu.com/s/1RJ3uyprMrqTrI2CNDstU6A?pwd=CGVQ (password:CGVQ)
