ExpoCM: Exposure-Aware One-Step Generative Single-Image HDR Reconstruction

2026-05-04

AI Summary

This work addresses the ill-posed challenge of high dynamic range (HDR) reconstruction from a single low dynamic range (LDR) image, where overexposed regions suffer from lost details and underexposed areas exhibit amplified noise. The authors propose an exposure-aware, single-step generative framework that formulates HDR reconstruction as a probability flow ordinary differential equation (PF-ODE). By leveraging soft exposure masks for spatial partitioning and region-conditioned consistent trajectories, the model jointly recovers fine details, suppresses noise, and preserves structural integrity in a single forward pass. A novel exposure-weighted luminance-chrominance loss in CIE L*a*b* color space enables high-quality reconstruction without requiring knowledge distillation. The method achieves state-of-the-art performance on HDR-REAL, HDR-EYE, and AIM2025 benchmarks, with inference speeds over 400× faster than DDPM (1000 steps) and 20× faster than DDIM (50 steps).

Abstract

Single-image HDR reconstruction aims to recover high dynamic range radiance from a single low dynamic range (LDR) input, but remains highly ill-posed due to detail saturation in over-exposed regions and noise amplification in under-exposed areas. While recent diffusion-based approaches offer powerful generative priors, they often overlook the exposure-dependent nature of the degradation and incur substantial computational costs from iterative sampling. To address these challenges, we propose ExpoCM, a novel one-step generative HDR reconstruction framework that reformulates HDR reconstruction as a Probability Flow ODE (PF-ODE) and constructs exposure-aware consistency trajectories via exposure-dependent perturbations. Specifically, a soft exposure mask is first constructed to separate the LDR image into over-, under-, and well-exposed regions. Based on this partition, region-conditioned consistency trajectories are designed to hallucinate saturated details, suppress noise in dark regions, and preserve reliable structures within a single, distillation-free inference step. To further enhance perceptual quality, we introduce an Exposure-guided Luminance-Chromaticity Loss in the CIE~$\text{L}^*\text{a}^*\text{b}^*$ space, which assigns exposure-aware weights to luminance and chromaticity components, effectively mitigating brightness bias and color drift. Extensive experiments on the HDR-REAL, HDR-EYE, and AIM2025 benchmarks demonstrate that ExpoCM achieves state-of-the-art fidelity and perceptual accuracy, while enabling over 400$\times$ and 20$\times$ faster inference compared to DDPM (1000 steps) and DDIM (50 steps), respectively.

Problem

• HDR reconstruction
• single-image
• exposure-aware
• ill-posed
• dynamic range

Innovation

• Exposure-aware
• One-step generation
• PF-ODE
• HDR reconstruction
• Luminance-Chromaticity Loss