Refine-IQA: Multi-Stage Reinforcement Finetuning for Perceptual Image Quality Assessment

TL;DR

Refine-IQA introduces a multi-stage reinforcement fine-tuning framework that enhances perceptual image quality assessment by explicitly improving the model's visual perception and interpretative capabilities, leading to superior performance.

Contribution

The paper proposes a novel multi-stage RFT framework with a new dataset and reward functions to improve low-level visual perception and interpretability in IQA models.

Findings

Achieves outstanding performance on perception and scoring tasks.

Activates robust 'think' (interpretation) capabilities in the model.

Excels on the quality interpreting benchmark.

Abstract

Reinforcement fine-tuning (RFT) is a proliferating paradigm for LMM training. Analogous to high-level reasoning tasks, RFT is similarly applicable to low-level vision domains, including image quality assessment (IQA). Existing RFT-based IQA methods typically use rule-based output rewards to verify the model's rollouts but provide no reward supervision for the "think" process, leaving its correctness and efficacy uncontrolled. Furthermore, these methods typically fine-tune directly on downstream IQA tasks without explicitly enhancing the model's native low-level visual quality perception, which may constrain its performance upper bound. In response to these gaps, we propose the multi-stage RFT IQA framework (Refine-IQA). In Stage-1, we build the Refine-Perception-20K dataset (with 12 main distortions, 20,907 locally-distorted images, and over 55K RFT samples) and design multi-task reward functions to strengthen the model's visual quality perception. In Stage-2, targeting the quality scoring task, we introduce a probability difference reward involved strategy for "think" process supervision. The resulting Refine-IQA Series Models achieve outstanding performance on both perception and scoring tasks-and, notably, our paradigm activates a robust "think" (quality interpreting) capability that also attains exceptional results on the corresponding quality interpreting benchmark.