Saved in:
Bibliographic Details
Main Authors: Xie, Wulin, Dai, Rui, Ding, Ruidong, Liu, Kaikui, Chu, Xiangxiang, Hou, Xinwen, Wen, Jie
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.22920
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914328706285568
author Xie, Wulin
Dai, Rui
Ding, Ruidong
Liu, Kaikui
Chu, Xiangxiang
Hou, Xinwen
Wen, Jie
author_facet Xie, Wulin
Dai, Rui
Ding, Ruidong
Liu, Kaikui
Chu, Xiangxiang
Hou, Xinwen
Wen, Jie
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2601_22920
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Q-Hawkeye: Reliable Visual Policy Optimization for Image Quality Assessment
Xie, Wulin
Dai, Rui
Ding, Ruidong
Liu, Kaikui
Chu, Xiangxiang
Hou, Xinwen
Wen, Jie
Computer Vision and Pattern Recognition
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.
title Q-Hawkeye: Reliable Visual Policy Optimization for Image Quality Assessment
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2601.22920