TIDE: Two-Stage Inverse Degradation Estimation with Guided Prior Disentanglement for Underwater Image Restoration

Underwater image restoration is essential for marine applications ranging from ecological monitoring to archaeological surveys, but effectively addressing the complex and spatially varying nature of underwater degradations remains a challenge. Existing methods typically apply uniform restoration strategies across the entire image, struggling to handle multiple co-occurring degradations that vary spatially and with water conditions. We introduce TIDE, a two stage inverse degradation estimation framework that explicitly models degradation characteristics and applies targeted restoration through specialized prior decomposition. Our approach disentangles the restoration process into multiple specialized hypotheses that are adaptively fused based on local degradation patterns, followed by a progressive refinement stage that corrects residual artifacts. Specifically, TIDE decomposes underwater degradations into four key factors, namely color distortion, haze, detail loss, and noise, and designs restoration experts specialized for each. By generating specialized restoration hypotheses, TIDE balances competing degradation factors and produces natural results even in highly degraded regions. Extensive experiments across both standard benchmarks and challenging turbid water conditions show that TIDE achieves competitive performance on reference based fidelity metrics while outperforming state of the art methods on non reference perceptual quality metrics, with strong improvements in color correction and contrast enhancement. Our code will be released upon acceptance.

Contributions

• We present TIDE, a framework that addresses complex, spatially varying underwater degradations through a structured, multi-stage restoration process.
• We introduce a degradation-specific hypothesis generation and fusion strategy, which produces targeted corrections for distinct degradation types and combines them adaptively to handle heterogeneous distortions.
• We develop a residual-aware refinement mechanism that selectively enhances poorly restored regions, improving overall fidelity and perceptual quality without compromising well-recovered areas.

We approach UIR via inverse degradation estimation and prior disentanglement. Instead of directly mapping degraded images to clean ones, we explicitly model degradations and apply targeted restoration. Different degradations require specialized treatment. Through prior disentanglement, we generate multiple hypotheses each addressing a specific degradation. TIDE implements this in two stages: first, degradation-guided multi-hypothesis restoration combines specialized hypotheses; second, a refinement stage corrects residual degradations.

The first stage of TIDE performs inverse degradation mapping to identify the spatial distribution and severity of different degradation types, followed by specialized prior decomposition to generate targeted restoration hypotheses. Our second stage implements a progressive refinement approach that explicitly identifies and addresses these remaining artifacts through differential degradation analysis and expert-guided correction.