Compressed Video Super-Resolution based on Hierarchical Encoding

2025-06-17

AI Summary

This work addresses the joint 4× super-resolution and artifact restoration for highly compressed H.265/HEVC video (e.g., upscaling from 180p to 720p or 270p to 1080p), where severe compression artifacts—including blocking, ringing, and blurring—impair perceptual quality. Method: We propose a hierarchical encoder Transformer architecture that explicitly models distortion characteristics across multiple quantization parameters (QPs) in the compressed domain. Leveraging multi-QP joint training and adaptive feature fusion, our method systematically suppresses diverse artifacts while preserving structural fidelity. Contribution/Results: The approach demonstrates strong cross-domain generalization—performing robustly on both generic video and talking-head sequences. It achieves state-of-the-art performance in both tracks of the ICME 2025 VSR Grand Challenge, significantly improving subjective visual quality and fine-detail preservation. This enables efficient real-time deployment in bandwidth-constrained applications such as video conferencing.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Learning on the Edge & Model Compression
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

This paper presents a general-purpose video super-resolution (VSR) method, dubbed VSR-HE, specifically designed to enhance the perceptual quality of compressed content. Targeting scenarios characterized by heavy compression, the method upscales low-resolution videos by a ratio of four, from 180p to 720p or from 270p to 1080p. VSR-HE adopts hierarchical encoding transformer blocks and has been sophisticatedly optimized to eliminate a wide range of compression artifacts commonly introduced by H.265/HEVC encoding across various quantization parameter (QP) levels. To ensure robustness and generalization, the model is trained and evaluated under diverse compression settings, allowing it to effectively restore fine-grained details and preserve visual fidelity. The proposed VSR-HE has been officially submitted to the ICME 2025 Grand Challenge on VSR for Video Conferencing (Team BVI-VSR), under both the Track 1 (General-Purpose Real-World Video Content) and Track 2 (Talking Head Videos).

Problem

• Enhance perceptual quality of compressed videos
• Upscale low-resolution videos by four times
• Eliminate compression artifacts from H.265/HEVC encoding

Innovation

• Hierarchical encoding transformer blocks for VSR
• Optimized to remove H.265/HEVC compression artifacts
• Upscales low-res videos by four times ratio