LoopLUT: 3D Lookup Tables with Progressive Region Refinement for Real-Time 4K Image Enhancement

2026-09-27

AI Summary

This study addresses the limitations of global 3D LUTs in handling spatially heterogeneous color enhancement and their computational constraints for real-time 4K processing. To this end, we propose a cascaded region-refined 3D LUT architecture. This method employs a low-resolution gating network to predict regions requiring correction and constructs residual LUTs for progressive optimization. By leveraging the partition of unity formed by the gates, convex combination fusion is performed directly, thereby avoiding interpolation error accumulation without requiring additional modules. Experimental results demonstrate that the proposed architecture achieves up to a 2.81 dB improvement in PSNR across multiple benchmark and underwater datasets while maintaining real-time throughput for 4K resolution, effectively balancing enhancement accuracy with computational efficiency.

Abstract

Color enhancement of 4K images must meet a quality target under a tight compute budget. Three-dimensional lookup tables (3D LUTs) dominate real-time enhancement because they decide at low resolution and apply a per-pixel lookup at full resolution. A single global LUT, however, is spatially invariant, so an underexposed shadow and a well-exposed region that share a pixel value receive identical corrections. Spatially heterogeneous demands cannot be expressed by such a mapping. We propose LoopLUT, a region-cascaded 3D LUT with progressive refinement. A global LUT performs the overall correction, followed by K-1 loop iterations. In each iteration a gating head predicts at low resolution the region that still needs correction, then builds a residual LUT from the color statistics of that region alone. The cascaded gates form a partition of unity, so the output is a per-pixel convex combination of the K lookup results. Fusion is therefore performed by the gates themselves, with no separate fusion module and no interpolation error accumulating across rounds. The decision stage runs at a fixed 256x256 resolution, independent of output resolution, so a 4K image costs only K pure lookups. Extensive experiments across four benchmarks show that LoopLUT improves PSNR by up to 2.81 dB over the strongest prior method, while keeping real-time throughput at 4K. The same decomposition also generalizes well to underwater enhancement datasets.

Problem

• Image Enhancement
• 3D Lookup Tables
• Real-time 4K Processing
• Spatial Heterogeneity
• Color Correction

Innovation

• 3D Lookup Tables
• Progressive Region Refinement
• Gating Mechanism
• Real-Time 4K Enhancement
• Residual LUT