Quality Consistency Score (QCS)

This repository provides a minimal reference implementation of the Quality Consistency Score (QCS), a post-processing temporal reliability descriptor computed from a per-frame (or per-second) quality signal (e.g., VMAF, PSNR, SSIM/MS-SSIM, or MOS-over-time).

Definition

Let $Q(t)$ be a time series of quality scores and let the survival curve be:

• $S(x) = P(Q \ge x)$ (probability scale), or
• $S(x) = 100 \cdot P(Q \ge x)$ (percent scale).

Then the Quality Consistency Score over an operating interval $[a,b]$ is:

$$ \mathrm{QCS}_{[a,b]} = \frac{1}{b-a} \int_a^b S(x),dx $$

qcs.py supports output in probability (scale="prob", range [0,1]) or percent (scale="percent", range [0,100]).

Repository contents

• qcs.py — QCS + survival curve + a few pooling baselines (arithmetic mean, harmonic mean, Minkowski p=0.5, 10th-percentile pooling, worst-1%)
• qcs_sample.ipynb — demo notebook using the sample traces
• VMAF_scores/ — sample per-frame VMAF traces

Quickstart

Requirements

pip install numpy matplotlib