GMODiff: One-Step Gain Map Refinement with Diffusion Priors for HDR Reconstruction
This work addresses HDR reconstruction for imaging applications, offering a faster and more efficient method, though it is incremental as it builds on existing LDM techniques with specific refinements.
The paper tackles the challenge of applying pre-trained Latent Diffusion Models (LDMs) to multi-exposure High Dynamic Range (HDR) reconstruction by introducing GMODiff, a one-step diffusion framework that estimates gain maps instead of full HDR content, achieving results comparable to state-of-the-art methods and being 100 times faster than previous LDM-based approaches.
Pre-trained Latent Diffusion Models (LDMs) have recently shown strong perceptual priors for low-level vision tasks, making them a promising direction for multi-exposure High Dynamic Range (HDR) reconstruction. However, directly applying LDMs to HDR remains challenging due to: (1) limited dynamic-range representation caused by 8-bit latent compression, (2) high inference cost from multi-step denoising, and (3) content hallucination inherent to generative nature. To address these challenges, we introduce GMODiff, a gain map-driven one-step diffusion framework for multi-exposure HDR reconstruction. Instead of reconstructing full HDR content, we reformulate HDR reconstruction as a conditionally guided Gain Map (GM) estimation task, where the GM encodes the extended dynamic range while retaining the same bit depth as LDR images. We initialize the denoising process from an informative regression-based estimate rather than pure noise, enabling the model to generate high-quality GMs in a single denoising step. Furthermore, recognizing that regression-based models excel in content fidelity while LDMs favor perceptual quality, we leverage regression priors to guide both the denoising process and latent decoding of the LDM, suppressing hallucinations while preserving structural accuracy. Extensive experiments demonstrate that our GMODiff performs favorably against several state-of-the-art methods and is 100 faster than previous LDM-based methods.