Closed Susmit-A closed 3 years ago
Extracts of the script which I use to compute all the scores can be found below. Note that I read the sequences with skvideo.io.vread as grayscale sequences. I basically use pandas to compute the means for all scores for each value of sigma and for each algorithm.
from skvideo.io import vread
from skvideo.measure import viideo_score, strred
...
# open the reference sequence
refseq = vread(seq_ref_path, num_frames=nframes, outputdict={"-pix_fmt": "gray"})
# iterate over the list of sigmas and algorithms
for sigma, algo in product(sigmaL, algoD.keys()):
denseq = vread(seq_den_path, num_frames=nframes, outputdict={"-pix_fmt": "gray"})
_, strred_fr, strred_rr = strred(refseq, denseq)
...
# save all the scores as csv
Later, I use pandas to open csv and compute the mean for each metric. The results are the values which appear in the paper.
How do you obtain the ST-RRED scores?
The scikit-video implementation returns 3 values -
(strred_array, strred, strredssn)
. On DAVIS, with sigma=10, all three values for denoised videos have a mean less than 0.1. The PSNR for the same set of videos is 38.94.The sum over
strred_array
on each video gives an average of5.0790
over all videos.