Image quality assessment improved with local and global visual analysis
VGIQA: Fine-Grained Image Quality Assessment with Visual Grounding Rewards
Computer Vision and Pattern Recognition
Summary
Evaluating how good a picture looks is tricky because quality can vary across different parts of the image. The authors developed a new method called VGIQA that looks at both the overall image and specific areas to judge image quality better. They trained their model in two steps and created a new dataset to help it focus on smaller details. Their method outperforms some existing large models and also helps detect image problems more accurately.
What this means in practice
- •For photo editing software developers: Integrate local and global image quality evaluation to suggest precise edits improving image clarity and appearance.
- •For social media platform engineers: Automatically detect and highlight areas of degraded quality in user-uploaded images for better content moderation and enhancement suggestions.
Authors
Weichang Wang, Liyu Chen, Shuyue Guan, Jie Hu, Fei Gao, Wei Li, Nannan Wang
Abstract
Image quality assessment (IQA) aims to evaluate the perceptual visual quality of images. Recently, the rapid development of Multi-modal Large Language Models (MLLMs) has significantly advanced the IQA field. By integrating chain-of-thought (CoT) reasoning and reinforcement learning (RL), these models can generate more accurate and interpretable results while relying on less training data. However, existing MLLM-based methods are predominantly constrained to holistic, image-level evaluations and lack the capability to produce fine-grained, localized assessments, which substantially limits their practical applicability. To address this limitation, we propose VGIQA, a novel IQA model endowed with both local and global quality assessment capabilities. VGIQA is trained via a two-stage paradigm comprising a cold-start phase followed by reinforcement learning. Driven by carefully designed local quality assessment reward functions, our model exhibits exceptional image quality understanding on VGSR-2K, a new dataset we propose specifically for local quality evaluation, significantly outperforming existing open-source MLLMs such as Qwen2.5-VL-72B-Instruct and Qwen3-VL-32B-Instruct. Furthermore, the local quality evaluation task can effectively enhance the model's performance on the degradation perception task.