ICME 2025 Grand Challenge on Video Super-Resolution for Video Conferencing

TL;DR

This paper introduces the ICME 2025 Grand Challenge on Video Super-Resolution for Video Conferencing, focusing on developing causal models to enhance low-resolution videos in real-time conferencing scenarios across multiple content types.

Contribution

It presents a new benchmark challenge with datasets and evaluation protocols for VSR in conferencing, including a novel screen content dataset and a focus on perceptual quality under low-delay constraints.

Findings

Open-sourced a new screen content dataset for VSR.

Evaluated models using crowdsourced subjective tests.

Addressed VSR for multiple video content types in conferencing.

Abstract

Super-Resolution (SR) is a critical task in computer vision, focusing on reconstructing high-resolution (HR) images from low-resolution (LR) inputs. The field has seen significant progress through various challenges, particularly in single-image SR. Video Super-Resolution (VSR) extends this to the temporal domain, aiming to enhance video quality using methods like local, uni-, bi-directional propagation, or traditional upscaling followed by restoration. This challenge addresses VSR for conferencing, where LR videos are encoded with H.265 at fixed QPs. The goal is to upscale videos by a specific factor, providing HR outputs with enhanced perceptual quality under a low-delay scenario using causal models. The challenge included three tracks: general-purpose videos, talking head videos, and screen content videos, with separate datasets provided by the organizers for training, validation, and testing. We open-sourced a new screen content dataset for the SR task in this challenge. Submissions were evaluated through subjective tests using a crowdsourced implementation of the ITU-T Rec P.910.