LiBrA-Net: Lie-Algebraic Bilateral Affine Fields for Real-Time 4K Video Dehazing

TL;DR

LiBrA-Net is a real-time 4K video dehazing method that uses Lie algebraic bilinear affine fields, coupled with a new benchmark dataset, to achieve state-of-the-art results efficiently.

Contribution

The paper introduces LiBrA-Net, a novel approach leveraging Lie algebraic group theory for efficient, high-quality 4K video dehazing, and releases the UHV-4K benchmark dataset.

Findings

LiBrA-Net runs at 25 FPS on 4K videos with only 6.12 million parameters.

It outperforms existing methods on UHV-4K, REVIDE, and HazeWorld datasets.

The UHV-4K benchmark provides comprehensive annotations for 4K video dehazing evaluation.

Abstract

Currently, there is a gap in the field of ultra-high-definition (UHD) video dehazing due to the lack of a benchmark for evaluation. Furthermore, existing video dehazing methods cannot run on consumer-grade GPUs when processing continuous UHD sequences of 3--5 frames at a time. In this paper, we address both issues with a new benchmark and an efficient method. Our key observation is that atmospheric dehazing reduces to a per-pixel affine transform governed by the low-frequency depth field, which can be compactly encoded in bilateral grids whose prediction cost is decoupled from the output resolution. Building on this, we propose LiBrA-Net, which factorizes the spatiotemporal affine field into a spatial--color and a temporal bilateral sub-grid predicted at a fixed low resolution, fuses their coefficients in the $\mathfrak{gl}(3)$ Lie algebra under group-theoretic regularization, maps the result to invertible GL(3) transforms via a Cayley parameterization, and restores high-frequency detail through a lightweight input-guided branch. We further release UHV-4K, the first paired 4K video dehazing benchmark with depth, transmission, and optical-flow annotations on every frame. Across UHV-4K, REVIDE, and HazeWorld, LiBrA-Net sets a new state of the art among compared video dehazing methods while running native 4K at 25 FPS on a single GPU with only 6.12 M parameters. Code and data are available at https://anonymous.4open.science/r/LiBrA-Net-42B8.