About the work
In this paper, we introduce Quality Consistency Score (QCS), a metric-agnostic temporal descriptor designed to complement traditional video quality evaluation methods.
Built on per-frame quality metrics such as PSNR, SSIM, and VMAF, QCS focuses on a critical but often overlooked aspect of video quality: how consistently quality is maintained over time.
Rather than relying solely on averages, QCS highlights transient quality drops—short-lived degradations that can significantly impact viewer experience but are often hidden in standard metrics like Rate–Distortion (RD) and BD-rate analysis.
Why it matters
Our results show that QCS provides valuable additional insight into codec performance by:
Revealing temporal instability in video quality
Complementing existing evaluation frameworks
Helping identify issues that affect perceived quality, even when average scores remain high
Looking ahead
We look forward to presenting this work at ICIP 2026 and engaging with the global research community.
Thanks to everyone involved for their contributions and support.
