VQ-Insight: Teaching VLMs for AI-Generated Video Quality Understanding via Progressive Visual Reinforcement Learning

TL;DR

VQ-Insight introduces a progressive, reasoning-style vision-language model that improves AI-generated video quality assessment by combining multi-stage learning and specialized rewards, outperforming existing methods.

Contribution

The paper presents VQ-Insight, a novel framework that integrates progressive learning and reward design to enhance video quality evaluation for AI-generated videos.

Findings

Outperforms state-of-the-art baselines in preference comparison.

Achieves superior multi-dimension scoring accuracy.

Enhances natural video scoring performance.

Abstract

Recent advances in AI-generated content (AIGC) have led to the emergence of powerful text-to-video generation models. Despite these successes, evaluating the quality of AIGC-generated videos remains challenging due to limited generalization, lack of temporal awareness, heavy reliance on large-scale annotated datasets, and the lack of effective interaction with generation models. Most current approaches rely on supervised finetuning of vision-language models (VLMs), which often require large-scale annotated datasets and tend to decouple understanding and generation. To address these shortcomings, we propose VQ-Insight, a novel reasoning-style VLM framework for AIGC video quality assessment. Our approach features: (1) a progressive video quality learning scheme that combines image quality warm-up, general task-specific temporal learning, and joint optimization with the video generation model; (2) the design of multi-dimension scoring rewards, preference comparison rewards, and temporal modeling rewards to enhance both generalization and specialization in video quality evaluation. Extensive experiments demonstrate that VQ-Insight consistently outperforms state-of-the-art baselines in preference comparison, multi-dimension scoring, and natural video scoring, bringing significant improvements for video generation tasks.