Grounding-IQA

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #typescript #llm #go #golang #android #machine-learning #rust #deep-learning #linux

★ 64 Open GitHub ↗

[ICLR'26] "Grounding-IQA: Grounding Multimodal Language Model for Image Quality Assessment"

Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Grounding-IQA: Grounding Multimodal Language Model for Image Quality Assessment

[Zheng Chen](https://zhengchen1999.github.io/), [Xun Zhang](https://scholar.google.com.hk/citations?hl=zh-CN&user=xCuW6LcAAAAJ), [Wenbo Li](https://fenglinglwb.github.io/), [Renjing Pei](https://orcid.org/0000-0001-7513-6576), [Fenglong Song](https://scholar.google.com/citations?hl=zh-CN&pli=1&user=WYDVk5oAAAAJ), [Xiongkuo Min](https://minxiongkuo.github.io/), [Xiaohong Liu](https://jhc.sjtu.edu.cn/~xiaohongliu/), [Xin Yuan](https://en.westlake.edu.cn/faculty/xin-yuan.html), [Yong Guo](https://www.guoyongcs.com/), and [Yulun Zhang](http://yulunzhang.com/), "Grounding-IQA: Grounding Multimodal Language Model for Image Quality Assessment", ICLR 2026

<div>
<a href="https://github.com/zhengchen1999/Grounding-IQA/releases" target='_blank' style="text-decoration: none;"><img src="https://img.shields.io/github/downloads/zhengchen1999/Grounding-IQA/total?color=green&style=flat"></a>
<a href="https://github.com/zhengchen1999/Grounding-IQA" target='_blank' style="text-decoration: none;"><img src="https://visitor-badge.laobi.icu/badge?page_id=zhengchen1999/Grounding-IQA"></a>
<a href="https://github.com/zhengchen1999/Grounding-IQA/stargazers" target='_blank' style="text-decoration: none;"><img src="https://img.shields.io/github/stars/zhengchen1999/Grounding-IQA?style=social"></a>
</div>
[[paper](https://openreview.net/pdf?id=yEpE0QPpf8)] [[project](https://zhengchen1999.github.io/Grounding-IQA)] [[arXiv](https://arxiv.org/abs/2411.17237)] [[supplementary material](https://github.com/zhengchen1999/Grounding-IQA/releases/download/v1/Supplementary_Material.pdf)] [dataset] [pretrained models]

#### 🔥🔥🔥 News

- **2026-01-26:** Grounding-IQA is accepted at ICLR 2026. 🎉🎉🎉
- **2024-11-26:** This repo is released.

---

> **Abstract:** The development of multimodal large language models (MLLMs) enables the evaluation of image quality through natural language descriptions. This advancement allows for more detailed assessments. However, these MLLM-based IQA methods primarily rely on general contextual descriptions, sometimes limiting fine-grained quality assessment.
> To address this limitation, we introduce a new image quality assessment (IQA) task paradigm, grounding-IQA. This paradigm integrates multimodal referring and grounding with IQA to realize more fine-grained quality perception. Specifically, grounding-IQA comprises two subtasks: grounding-IQA-description (GIQA-DES) and visual question answering (GIQA-VQA). GIQA-DES involves detailed descriptions with precise locations (e.g., bounding boxes), while GIQA-VQA focuses on quality QA for local regions. To realize grounding-IQA, we construct a corresponding dataset, GIQA-160K, through our proposed automated annotation pipeline. Furthermore, we develop a well-designed benchmark, GIQA-Bench. The benchmark comprehensively evaluates the model grounding-IQA performance from three perspectives: description quality, VQA accuracy, and grounding precision. Experiments demonstrate that our proposed task paradigm, dataset, and benchmark facilitate the more fine-grained IQA application.

![](https://raw.githubusercontent.com/zhengchen1999/Grounding-IQA/main/figs/Example.png)

---

### Pipeline

![](https://raw.githubusercontent.com/zhengchen1999/Grounding-IQA/main/figs/Pipeline.png)

---

### Radar Chat

![](https://raw.githubusercontent.com/zhengchen1999/Grounding-IQA/main/figs/Radar-Chat.png)

## 🔖 TODO

- [ ] Release testing and training code.
- [ ] Release GIQA-160K and GIQA-Bench.
- [ ] Release pre-trained models.
- [ ] Provide WebUI.
- [ ] Provide HuggingFace demo.

## 🔗 Contents

1. [Datasets](#datasets)
1. Models
1. Training
1. Testing
1. [Results](#results)
1. [Acknowledgements](#acknowledgements)

## <a name="datasets"></a>📦 Datasets

![](https://raw.githubusercontent.com/zhengchen1999/Grounding-IQA/main/figs/Dataset.png)

## <a name="results"></a>🔎 Results

We achieve impressive performance on GIQA-DES and GIQA-VQA tasks.

<details open>
<summary>Quantitative Results (click to expand)</summary>

- Results in Tab. 5 of the main paper

<details open>
<summary>Qualitative Results (click to expand)</summary>

- Results in Fig. 7 of the main paper

<p align="center">
  <img width="900" src="figs/Qualitative-1.png">
</p>
<details>
<summary>More Qualitative Results</summary>

- More Results on GIQA-DES (Fig. 6 of the supplementary material)

- More Results on GIQA-VQA (Fig. 7 of the supplementary material)

</details>

## <a name="citation"></a>📎 Citation

If you find the code helpful in your research or work, please cite the following paper(s).

```
@inproceedings{chen2026grounding,
  title={Grounding-IQA: Grounding Multimodal Language Model for Image Quality Assessment},
  author={Chen, Zheng and Zhang, Xun and Li, Wenbo and Pei, Renjing and Song, Fenglong and Min, Xiongkuo and Liu, Xiaohong and Yuan, Xin and Guo, Yong and Zhang, Yulun},
  booktitle={ICLR},
  year={2026}
}
```

## <a name="acknowledgements"></a>💡 Acknowledgements

This project is based on [Q-Instruct](https://github.com/Q-Future/Q-Instruct), [DepictQA](https://github.com/XPixelGroup/DepictQA), [mPLUG-Owl](https://github.com/X-PLUG/mPLUG-Owl), and [LLaVA](https://github.com/haotian-liu/LLaVA).

Grounding-IQA

Download

Today's picks

Was this content helpful?

Recent Feedback

Grounding-IQA

Download

Today's picks

Was this content helpful?

Recent Feedback

Download README