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

2025-06-13

AI Summary

This work addresses low-latency causal video super-resolution (VSR) under H.265 compression for video conferencing, targeting perceptual quality enhancement of three LR video types: generic videos, portrait close-ups, and screen content. To this end, we propose a causal temporal propagation network, a lightweight spatio-temporal feature alignment mechanism, and an H.265 distortion modeling approach. We further introduce the first open-source, conference-oriented screen-content VSR dataset. Subjective evaluation follows ITU-T P.910 guidelines via crowdsourcing. Experiments demonstrate that our method—and top-performing competition solutions—significantly outperform bilinear interpolation and non-causal VSR baselines under strict low-latency constraints. The results advance the practical deployment of real-time VSR under realistic codec conditions.

Technology Category

Computer Vision: Image and Video RetrievalMachine Learning: Causal LearningKnowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingResponsible Web: Human-perceived consequences of algorithmic deployment on the webGraph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs

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.

Problem

• Enhance video quality for conferencing using super-resolution
• Upscale low-resolution videos with H.265 encoding
• Evaluate methods on diverse video tracks and datasets

Innovation

• Upscale videos using causal models
• H.265 encoded LR video enhancement
• Open-sourced screen content dataset