DY-LUT: Depth-Aware YCbCr Lookup Tables for Real-Time Underwater Image Enhancement

2026-07-24

AI Summary

This work addresses the spatially non-uniform degradation of underwater images caused by wavelength-dependent light attenuation over varying propagation distances. To tackle this challenge, the authors propose the DY-LUT framework, which introduces depth-awareness into the YCbCr color space for the first time. A dual-branch encoder predicts both fusion weights and pixel-wise degradation indices, which jointly modulate a learnable 4D lookup table to enable efficient, spatially adaptive, and real-time enhancement. Evaluated on the UIEB-90 and LSUI datasets, the method achieves competitive visual quality, processes 4K images in approximately 7 milliseconds—9 to 304 times faster than state-of-the-art large models—and significantly improves downstream tasks such as object detection and feature matching, thereby demonstrating the superiority of YCbCr as a foundational representation for conditional lookup tables.

Abstract

Underwater image enhancement is challenged by spatially non-uniform, wavelength-dependent attenuation. Propagation distance and wavelength govern this degradation, while YCbCr separates luminance from chrominance for restoration. We propose DY-LUT, a depth-aware YCbCr lookup-table framework for real-time enhancement. A dual-branch encoder predicts image-level fusion weights and a joint pair of pixel-wise degradation indices from image and depth features. These quantities condition learnable 4D LUTs, followed by lightweight local refinement. DY-LUT preserves traditional LUT efficiency while enabling depth-conditioned, spatially adaptive restoration. With externally supplied depth, its 3.56M-parameter enhancement network achieves competitive quality on UIEB-90 and LSUI and runs $9$--$304\times$ faster than representative high-capacity baselines. Adaptive inference further maintains real-time performance ($\sim7$ ms) for 4K UIQAD images. DY-LUT also benefits downstream detection and feature matching. Ablations show that YCbCr is a more effective basis than RGB for depth-conditioned lookup, while the jointly learned indices further improve adaptive querying. These results provide a physically grounded route to efficient UIE on practical platforms.

Problem

• underwater image enhancement
• wavelength-dependent attenuation
• spatially non-uniform degradation
• depth-aware restoration

Innovation

• Depth-aware
• YCbCr
• Lookup Table (LUT)
• Real-time enhancement
• Underwater image restoration