IQA-T1: Tool-based Visual Evidence Reasoning for Image Quality Assessment

2026-07-14

AI Summary

This work addresses the challenges of weak generalization and poor interpretability in open-world image quality assessment, where existing multimodal large language model (MLLM)-based approaches are prone to semantic bias and struggle to capture low-level visual degradations. To overcome these limitations, the study introduces a tool-calling mechanism into the assessment pipeline, proposing a vision-evidence-based reasoning framework. During inference, the model autonomously invokes specialized analytical tools—such as noise residual estimation, gradient statistics, and spectral analysis—to generate structured visual evidence, which is then integrated into the MLLM to construct an interpretable reasoning chain. The authors release Q-Tool, a dataset comprising 11k multimodal reasoning chains, and demonstrate state-of-the-art performance across seven benchmarks, significantly improving both assessment accuracy and model interpretability.

Abstract

Image Quality Assessment (IQA) in open-world environments remains challenging due to limited generalization and interpretability. Recent approaches based on multimodal large language models (MLLMs) introduce textual reasoning for quality prediction, yet their judgments rely heavily on semantically biased internal representations, making them insensitive to low-level perceptual degradations. We propose IQA-T1, a tool-based visual evidence reasoning framework that augments MLLM reasoning with explicit perceptual observations. During inference, the model autonomously invokes specialized analysis tools to generate structured visual evidence, such as noise residual maps, gradient statistics, and frequency spectra, which are progressively integrated into the reasoning process. To support this paradigm, we construct Q-Tool, a dataset containing 11k multimodal reasoning chains grounded in tool-generated evidence. Extensive experiments on seven IQA benchmarks show that IQA-T1 achieves the best overall performance across datasets while producing interpretable and evidence-grounded quality assessments. Code and dataset are available at https://github.com/zibuyu-02/IQA-T1.

Problem

• Image Quality Assessment
• open-world generalization
• perceptual degradations
• interpretability
• multimodal large language models

Innovation

• tool-based reasoning
• visual evidence
• image quality assessment
• multimodal LLM
• perceptual degradation