Comparison Drives Preference: Reference-Aware Modeling for AI-Generated Video Quality Assessment

2026-04-18

AI Summary

This work addresses the limitation of existing AI-generated video quality assessment methods, which often neglect inter-video relationships and fail to emulate the human comparative perception mechanism. To overcome this, the authors propose a reference-aware evaluation paradigm that, for the first time, introduces graph structures into this task. Specifically, a reference graph is constructed centered on the query video, and a graph-guided difference aggregation mechanism is employed to integrate information from related videos. This is coupled with a multi-dimensional quality prediction model to yield the final assessment. The proposed method outperforms state-of-the-art approaches across multiple datasets and demonstrates strong generalization capability in cross-dataset settings, thereby validating the efficacy of explicitly modeling inter-video relationships for improving assessment accuracy.

Technology Category

• Computer Vision: Video Understanding & Activity Analysis
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Knowledge Representation and Reasoning: Qualitative Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

The rapid advancement of generative models has led to a growing volume of AI-generated videos, making the automatic quality assessment of such videos increasingly important. Existing AI-generated content video quality assessment (AIGC-VQA) methods typically estimate visual quality by analyzing each video independently, ignoring potential relationships among videos. In this work, we revisit AIGC-VQA from an inter-video perspective and formulate it as a reference-aware evaluation problem. Through this formulation, quality assessment is guided not only by intrinsic video characteristics but also by comparisons with related videos, which is more consistent with human perception. To validate its effectiveness, we propose Reference-aware Video Quality Assessment (RefVQA), which utilizes a query-centered reference graph to organize semantically related samples and performs graph-guided difference aggregation from the reference nodes to the query node. Experiments on existing datasets demonstrate that our proposed RefVQA outperforms state-of-the-art methods across multiple quality dimensions, with strong generalization ability validated by cross-dataset evaluation. These results highlight the effectiveness of the proposed reference-based formulation and suggest its potential to advance AIGC-VQA.

Problem

• AI-generated video
• quality assessment
• reference-aware
• inter-video comparison
• AIGC-VQA

Innovation

• reference-aware modeling
• inter-video comparison
• graph-guided aggregation
• AIGC-VQA
• query-centered reference graph