Expert Operational GANS: Towards Real-Color Underwater Image Restoration

2025-07-14

AI Summary

Underwater image restoration suffers from severe heterogeneous degradation caused by light scattering, absorption, and depth-dependent attenuation, which challenges single-generator GANs in modeling the full spectrum of degradation patterns. To address this, we propose xOp-GAN—a multi-expert generative adversarial network—featuring multiple specialized generators collaboratively handling images with varying degradation severities. Crucially, we introduce the first discriminative confidence-guided adaptive selection mechanism at inference time, leveraging the discriminator’s perceptual confidence scores to dynamically choose the optimal restoration output. Additionally, we design a depth-hierarchical training strategy and a joint PSNR-perception optimization objective. Evaluated on the LSUI dataset, xOp-GAN achieves a PSNR of 25.16 dB, significantly outperforming single-generator baselines while reducing both model parameters and computational overhead. Our core contributions lie in (1) the discriminative-confidence-driven adaptive image selection mechanism and (2) the multi-expert collaborative architecture explicitly tailored for heterogeneous underwater degradations.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

The wide range of deformation artifacts that arise from complex light propagation, scattering, and depth-dependent attenuation makes the underwater image restoration to remain a challenging problem. Like other single deep regressor networks, conventional GAN-based restoration methods struggle to perform well across this heterogeneous domain, since a single generator network is typically insufficient to capture the full range of visual degradations. In order to overcome this limitation, we propose xOp-GAN, a novel GAN model with several expert generator networks, each trained solely on a particular subset with a certain image quality. Thus, each generator can learn to maximize its restoration performance for a particular quality range. Once a xOp-GAN is trained, each generator can restore the input image and the best restored image can then be selected by the discriminator based on its perceptual confidence score. As a result, xOP-GAN is the first GAN model with multiple generators where the discriminator is being used during the inference of the regression task. Experimental results on benchmark Large Scale Underwater Image (LSUI) dataset demonstrates that xOp-GAN achieves PSNR levels up to 25.16 dB, surpassing all single-regressor models by a large margin even, with reduced complexity.

Problem

• Restoring real-color underwater images with diverse degradations
• Overcoming limitations of single-generator GANs for heterogeneous domains
• Selecting optimal expert-generated restoration using discriminator confidence

Innovation

• Multiple expert generator networks for restoration
• Discriminator selects best image during inference
• Specialized training on subsets for quality ranges