Towards Distortion-Debiased Blind Image Quality Assessment

Lize Zhou, Xiaoqi Wang, Jian Xiong, Xianzhong Long, Hao Gao

Abstract: Existing blind image quality assessment (BIQA) models are susceptible to biases related to distortion intensity and domain. Intensity bias manifests as an over-sensitivity to severe distortions and under-estimation of minor ones, while domain bias stems from the discrepancies between synthetic and authentic distortion properties. This work introduces a unified learning framework to address these distortion biases. We integrate distortion perception and restoration modules to address intensity bias. The restoration module uses a combined image-level and feature-level denoising method to restore distorted images, where easily restorable minor distortions serve as references for mildly distorted images, and severe distortions benefit directly from distortion perception. Finally, calculating a distortion intensity matrix via intensity-aware cross-attention for adaptive handling of intensity bias. To tackle domain bias, we introduce a distortion domain recognition task, leveraging inherent differences between synthetic and authentic distortions for adaptive quality score weighting. Experimental results show that our proposed method achieves state-of-the-art performance on a multitude of synthetic and authentic IQA benchmark datasets. The code and models will be available.

Primary Subject Area: [Experience] Interactions and Quality of Experience

Secondary Subject Area: [Experience] Multimedia Applications

Relevance To Conference: With the widespread use of digital images in areas such as social media, online advertising, medical imaging, and remote monitoring, ensuring and controlling image quality becomes a critical task. However, existing blind image quality assessment (BIQA) models suffer from intensity bias and distortion domain bias. This work proposes a distortion-debiased BIQA method to address these biases. Extensive experiments demonstrate that the proposed approach can better handle diverse distortion scenarios, achieving state-of-the-art performance.

Submission Number: 5734