LUCID: Learning Unified Control for Image Deflaring and Exposure Mastery in Nighttime Photography

2026-06-05

AI Summary

Nighttime photography is often degraded by the coupled effects of intense glare-induced structural occlusion and low-light noise, which existing methods struggle to address jointly. This work proposes a unified controllable diffusion-based generative framework that disentangles glare through a dedicated module to extract structural guidance, and integrates classifier-free guidance (CFG) with continuous exposure control to enable fine-grained manipulation of light sources, glare, and HDR exposure. The method introduces a novel four-mode training strategy, formulating image restoration as a continuously controllable process. Evaluated on real-world nighttime scenes, it significantly outperforms state-of-the-art approaches while simultaneously achieving high-quality glare removal and high dynamic range reconstruction.

Abstract

Photography is the art of painting with light, yet nighttime scenes are shaped by competing degradations: intense flares obscure scene structure, while photon-limited regions collapse into noise. Conventional approaches address these factors in isolation, overlooking the fact that these degradations are fundamentally entangled. To bridge this gap, we introduce LUCID, a unified framework that reframes nighttime restoration as a continuous and controllable process rather than a fixed correction. We decompose nighttime restoration into two cooperative components: a flare disentanglement module that lifts the 'curtain' of optical artifacts to provide reliable structural guidance, and a diffusion-driven module that leverages generative priors to reconstruct clean and well-exposed imagery. Crucially, LUCID introduces explicit controllability through a novel four-mode training strategy, enabling users to steer the restoration process via classifier-free guidance (CFG) and allowing selective control over light sources and their associated flare and ghosting artifacts, while also supporting high dynamic range (HDR) reconstruction through continuous exposure control. Extensive experiments demonstrate that LUCID consistently outperforms state-of-the-art methods across diverse real-world nighttime scenarios.

Problem

• nighttime photography
• flare removal
• exposure control
• image restoration
• degradation entanglement

Innovation

• unified restoration
• flare disentanglement
• diffusion model
• controllable generation
• nighttime photography