In a recent face-to-face VQEG meeting (Dec 2022, online) it was agreed to move the activities of the VQEG IGVQM (Implementer’s Guide to Video Quality Metrics) group into JEG-Hybrid.

Originally, IGVQM had a quite broad scope as detailed in the original working doc, for instance:

  • Priority to full-reference (pixel) metrics, but also applicability to no-reference metrics
    • Use at least all metrics in VMAF library, focus on open-source solutions
  • Offer temporal aggregation methods of image quality metrics into video ones
  • Perform statistical analysis of existing subjective datasets (limited to compression and scaling)
  • Offer standard logistic mappings of objective metrics to normalized linear scale

Currently, in JEG-Hybrid, due to the limited resources, initially we plan to:

  • Restrict the scope (at least initially) by focusing on some specific aspects (e.g., temporal pooling)
  • Start with the already available objective measures in some datasets (e.g., HDTV, ITS4S)
  • Try to create synergies between the current JEG-Hybrid activities and the IGVQM ones
  • Include the existing JEG-Hybrid experimentations on AI observers while carrying out some of the activities envisaged by IGVMQ, e.g., objective metrics computation and analysis

Old IGVQM page on VQEG website: IGVQM
Also, it should be noted that IGVQM received in the past a pointer to ITU-T Technical paper HSTP-VID-WPOM titled “Working practices using objective metrics for evaluation of video coding efficiency experiments” (persistent link), to be taken into account in the activities