CompressedVQA-AEV: Full-Reference and No-Reference Quality Assessment Models for Asymmetric Encoded Videos

2026-07-05

AI Summary

This study addresses the lack of effective quality assessment methods for asymmetrically coded videos by proposing a dual-model framework. The full-reference model leverages Swin-B to extract multi-stage similarity statistics between reference and distorted videos, while the no-reference model integrates SigLIP2 and Swin-B for frame-level feature encoding, followed by temporal average pooling and cross-fold ensemble prediction of perceptual quality. This work is the first to jointly employ Swin-B and SigLIP2 for this task, introducing innovative multi-stage similarity modeling and cross-fold ensemble strategies. In the QoMEX 2026 Grand Challenge, the proposed full-reference model achieved first place, and the no-reference model ranked fourth.

Abstract

This report presents our solutions to the QoMEX 2026 Grand Challenge on Video Quality Assessment for Asymmetric Encoded Videos, comprising a full-reference (FR) model, CompressedVQA-AEV-FR, and a no-reference (NR) model, CompressedVQA-AEV-NR. The FR approach leverages a Swin-B backbone to extract multi-stage similarity statistics between reference and distorted videos for quality prediction. For the NR setting, our model employs complementary frame-level encoders based on SigLIP2 and Swin-B, followed by temporal mean pooling and cross-fold ensembling to estimate perceptual quality without reference data. Our CompressedVQA-AEV-FR achieves first place in the FR track of QoMEX 2026 Grand Challenge, while CompressedVQA-AEV-NR secures fourth place in the NR track, demonstrating the effectiveness of our proposed models. The code is available at https://github.com/sunwei925/CompressedVQA-AEV.

Problem

• Video Quality Assessment
• Asymmetric Encoded Videos
• Full-Reference
• No-Reference

Innovation

• asymmetric encoded videos
• full-reference VQA
• no-reference VQA
• Swin-B backbone
• SigLIP2 encoder