Q-Hawkeye: Reliable Visual Policy Optimization for Image Quality Assessment

TL;DR

Q-Hawkeye introduces a reliable RL-based framework for image quality assessment that uses uncertainty estimation and perception grounding to improve stability and accuracy over existing methods.

Contribution

It proposes a novel uncertainty-aware and perception-grounded optimization framework for RL-based IQA, addressing stability and perceptual reliability issues.

Findings

Outperforms state-of-the-art IQA methods in experiments.

Demonstrates better generalization across multiple datasets.

Stabilizes policy optimization through uncertainty reweighting.

Abstract

Image Quality Assessment (IQA) predicts perceptual quality scores consistent with human judgments. Recent RL-based IQA methods built on MLLMs focus on generating visual quality descriptions and scores, ignoring two key reliability limitations: (i) although the model's prediction stability varies significantly across training samples, existing GRPO-based methods apply uniform advantage weighting, thereby amplifying noisy signals from unstable samples in gradient updates; (ii) most works emphasize text-grounded reasoning over images while overlooking the model's visual perception ability of image content. In this paper, we propose Q-Hawkeye, an RL-based reliable visual policy optimization framework that redesigns the learning signal through unified Uncertainty-Aware Dynamic Optimization and Perception-Aware Optimization. Q-Hawkeye estimates predictive uncertainty using the variance of predicted scores across multiple rollouts and leverages this uncertainty to reweight each sample's update strength, stabilizing policy optimization. To strengthen perceptual reliability, we construct paired inputs of degraded images and their original images and introduce an Implicit Perception Loss that constrains the model to ground its quality judgments in genuine visual evidence. Extensive experiments demonstrate that Q-Hawkeye outperforms state-of-the-art methods and generalizes better across multiple datasets. Our dataset and code are available at https://github.com/AMAP-ML/Q-Hawkeye.