VIFP Video Quality Metric – Python Script

VIFP (Visual Information Fidelity in Pixel Domain) is a video quality metric, in the following papers you can find details on VIFP:

“A New Method for Color Image Quality Assessment“, by Niveditta Thakur et al., 2011

“Image Information and Visual Quality“,  Hamid Rahim Sheikh, Member, IEEE, Alan C. Bovik, Fellow, IEEE

In the python script vifp.py, which was modified from vifp.py from https://github.com/aizvorski/video-quality.git and verified against VQMT ,  VIFP scores are computed:

 

Usage:

–ref         reference (pristine) yuv420p video sequence
–dec        decoded yuv420p video sequence
–height   frame height (default 1080)
–width    frame width (default 1920)
-n              number of frames to process, if 0 then all, (default 0)

 

Example [compute VIFP scores of first 10 frames]:

python vifp.py –ref crowd_200frames_50fps.yuv –dec crowd_encoded_200frames.yuv –height 1080 –width 1920 -n 10

Frame=0 VIFP=0.835
Frame=1 VIFP=0.637
Frame=2 VIFP=0.639
Frame=3 VIFP=0.784
Frame=4 VIFP=0.597
Frame=5 VIFP=0.595
Frame=6 VIFP=0.685
Frame=7 VIFP=0.546
Frame=8 VIFP=0.539
Frame=9 VIFP=0.582

 

Note:

There is another source of python scripts to compute VIFP, SSIM, NIQE etc.