Abstract:Generally, humans are more skilled at perceiving differences between high-quality (HQ) and low-quality (LQ) images than directly judging the quality of a single LQ image. This situation also applies to image quality assessment (IQA). Although recent no-reference (NR-IQA) methods have made great progress to predict image quality free from the reference image, they still have the potential to achieve better performance since HQ image information is not fully exploited. In contrast, full-reference (FR-IQA) methods tend to provide more reliable quality evaluation, but its practicability is affected by the requirement for pixel-level aligned reference images. To address this, we firstly propose the content-variant reference method via knowledge distillation (CVRKD-IQA). Specifically, we use non-aligned reference (NAR) images to introduce various prior distributions of high-quality images. The comparisons of distribution differences between HQ and LQ images can help our model better assess the image quality. Further, the knowledge distillation transfers more HQ-LQ distribution difference information from the FR-teacher to the NAR-student and stabilizing CVRKD-IQA performance. Moreover, to fully mine the local-global combined information, while achieving faster inference speed, our model directly processes multiple image patches from the input with the MLP-mixer. Cross-dataset experiments verify that our model can outperform all NAR/NR-IQA SOTAs, even reach comparable performance with FR-IQA methods on some occasions. Since the content-variant and non-aligned reference HQ images are easy to obtain, our model can support more IQA applications with its relative robustness to content variations. Our code and more detailed elaborations of supplements are available: <a class="link-external link-https" href="https://github.com/guanghaoyin/CVRKD-IQA" rel="external noopener nofollow">this https URL</a>.

Causal-IQA: Towards the Generalization of Image Quality Assessment Based on Causal Inference

Image Quality Assessment: Integrating Model-Centric and Data-Centric Approaches

Causal Perception Inspired Representation Learning for Trustworthy Image Quality Assessment

Blind Image Quality Assessment With Active Inference

Regression-free Blind Image Quality Assessment with Content-Distortion Consistency

Blind Image Quality Assessment Via Cross-View Consistency

A survey on IQA

Blind Quality Assessment for in-the-Wild Images via Hierarchical Feature Fusion and Iterative Mixed Database Training

Subjective Opinions Matter: Controllable Image Quality Assessment Using Pseudo Reference Images

Learning Conditional Knowledge Distillation for Degraded-Reference Image Quality Assessment

Quantifying Visual Image Quality: A Bayesian View

TOPIQ: A Top-down Approach from Semantics to Distortions for Image Quality Assessment

Multiple hallucinated deep network for image quality assessment

From Distortion Manifold to Perceptual Quality: a Data Efficient Blind Image Quality Assessment Approach

Learning from Synthetic Data for Opinion-free Blind Image Quality Assessment in the Wild

ConIQA: A deep learning method for perceptual image quality assessment with limited data

Content-Variant Reference Image Quality Assessment via Knowledge Distillation

Hierarchical Curriculum Learning for No-Reference Image Quality Assessment

Learn to Evaluate Image Perceptual Quality Blindly from Statistics of Self-similarity

Cross-IQA: Unsupervised Learning for Image Quality Assessment

Learning to Rank for Blind Image Quality Assessment