Determine an effective restorer for blur-based degradation

Determine a restoration method that can reliably recover video regions degraded by the PRESLEY blur transport and improve delivered quality over the transmitted blurred video.

Background

PRESLEY evaluates several restoration backbones for different degradation families. For the blur transport, the tested methods include InstantIR, NAFNet, and an unsharp filtering control. The reported experiments do not show a meaningful advantage for the learned restorers over the trivial control or over the transmitted video.

Consequently, the restoration problem for blurred blocks remains unresolved: a suitable model or restoration strategy must exploit the blurred transmitted prior sufficiently to improve the quality received by the viewer.

References

On the blur transport InstantIR does not beat a trivial unsharp control costing ${\approx}12$\,s against ${\approx}774$--$837$\,s, and neither it nor NAFNet delivers meaningful gain against the transmitted video, so that transport's restorer question is open.

— Perceptual Refinement of an End-to-End Video Streaming Pipeline via Generative AI Layers  (2609.19215 - Artioli et al., 16 Sep 2026) in Section 4, subsection “Restoration and Regime Results” (Section 4.2)