PIGUIQA: A Physical Imaging Guided Perceptual Framework for Underwater Image Quality Assessment

2024-12-20

AI Summary

This paper addresses the inconsistency between physical distortions and human perception in underwater image quality assessment (UIQA). It is the first to model the joint physical effects of direct transmission attenuation and backscattering as the primary cause of perceptual degradation. We propose a physics-guided dual-path evaluation framework: a local path employs neighborhood attention to adaptively perceive distortions, while a global path jointly aggregates scene semantics and underwater distortion features. The framework integrates radiative transfer modeling, multi-scale feature fusion, and physics-constrained loss functions to ensure interpretability and perceptual consistency. Evaluated on UIEB, LSUI, and EUVP benchmarks, our method achieves state-of-the-art performance, improving average Spearman rank-order correlation coefficient (SROCC) by 3.2% and reducing cross-dataset generalization error by 27%. The source code is publicly available.

Technology Category

• Computer Vision: Low Level & Physics-based Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Uncertainty Representations

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization

Abstract

In this paper, we propose a Physical Imaging Guided perceptual framework for Underwater Image Quality Assessment (UIQA), termed PIGUIQA. First, we formulate UIQA as a comprehensive problem that considers the combined effects of direct transmission attenuation and backward scattering on image perception. By leveraging underwater radiative transfer theory, we systematically integrate physics-based imaging estimations to establish quantitative metrics for these distortions. Second, recognizing spatial variations in image content significance and human perceptual sensitivity to distortions, we design a module built upon a neighborhood attention mechanism for local perception of images. This module effectively captures subtle features in images, thereby enhancing the adaptive perception of distortions on the basis of local information. Third, by employing a global perceptual aggregator that further integrates holistic image scene with underwater distortion information, the proposed model accurately predicts image quality scores. Extensive experiments across multiple benchmarks demonstrate that PIGUIQA achieves state-of-the-art performance while maintaining robust cross-dataset generalizability. The implementation is publicly available at https://anonymous.4open.science/r/PIGUIQA-A465/

Problem

• Develops a framework for underwater image quality assessment.
• Integrates physics-based imaging to quantify distortion effects.
• Enhances distortion perception using local and global image features.

Innovation

• Integrates underwater radiative transfer theory
• Uses neighborhood attention for local perception
• Employs global perceptual aggregator for quality prediction