GMODiff: One-Step Gain Map Refinement with Diffusion Priors for HDR Reconstruction

• URL: http://arxiv.org/abs/2512.16357v1
• Date: Thu, 18 Dec 2025 09:50:25 GMT
• Title: GMODiff: One-Step Gain Map Refinement with Diffusion Priors for HDR Reconstruction
• Authors: Tao Hu, Weiyu Zhou, Yanjie Tu, Peng Wu, Wei Dong, Qingsen Yan, Yanning Zhang,
• Abstract summary: We introduce GMODiff, a gain map-driven one-step diffusion framework for multi-exposure HDR reconstruction.<n>Our GMO performs favorably against several state-of-the-art methods and is 100 faster than previous LDM-based methods.
• Score: 48.881484713994496
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: 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.

Related papers

• Deep Lightweight Unrolled Network for High Dynamic Range Modulo Imaging [19.49437461280304]
  Modulo-Imaging (MI) offers a promising alternative for expanding the dynamic dynamic range images by resetting the signal intensity when it reaches the intensity level.<n>We introduce the Scaling Equi term that facilitates self-tuning, thereby enabling the model to adapt to new images outside the original distribution.
  arXiv  Detail & Related papers  (2026-01-18T18:22:38Z)
• AP-LDM: Attentive and Progressive Latent Diffusion Model for Training-Free High-Resolution Image Generation [12.564266865237343]
  Latent diffusion models (LDMs) often experience significant structural distortions when directly generating high-resolution (HR) images. We propose an Attentive and Progressive LDM (AP-LDM) aimed at enhancing HR image quality while accelerating the generation process. AP-LDM decomposes the denoising process of LDMs into two stages: (i) attentive training-resolution denoising, and (ii) progressive high-resolution denoising.
  arXiv  Detail & Related papers  (2024-10-08T13:56:28Z)
• Bracket Diffusion: HDR Image Generation by Consistent LDR Denoising [29.45922922270381]
  We demonstrate generating HDR images using the concerted action of multiple black-box, pre-trained LDR image diffusion models.<n>We operate multiple denoising processes to generate multiple LDR brackets that together form a valid HDR result.<n>We demonstrate state-of-the-art unconditional and conditional restoration-type (LDR2) generative modeling results, yet in HDR.
  arXiv  Detail & Related papers  (2024-05-23T08:24:22Z)
• Generating Content for HDR Deghosting from Frequency View [56.103761824603644]
  Recent Diffusion Models (DMs) have been introduced in HDR imaging field. DMs require extensive iterations with large models to estimate entire images. We propose the Low-Frequency aware Diffusion (LF-Diff) model for ghost-free HDR imaging.
  arXiv  Detail & Related papers  (2024-04-01T01:32:11Z)
• Towards High-quality HDR Deghosting with Conditional Diffusion Models [88.83729417524823]
  High Dynamic Range (LDR) images can be recovered from several Low Dynamic Range (LDR) images by existing Deep Neural Networks (DNNs) techniques. DNNs still generate ghosting artifacts when LDR images have saturation and large motion. We formulate the HDR deghosting problem as an image generation that leverages LDR features as the diffusion model's condition.
  arXiv  Detail & Related papers  (2023-11-02T01:53:55Z)
• SMAE: Few-shot Learning for HDR Deghosting with Saturation-Aware Masked Autoencoders [97.64072440883392]
  We propose a novel semi-supervised approach to realize few-shot HDR imaging via two stages of training, called SSHDR. Unlikely previous methods, directly recovering content and removing ghosts simultaneously, which is hard to achieve optimum. Experiments demonstrate that SSHDR outperforms state-of-the-art methods quantitatively and qualitatively within and across different datasets.
  arXiv  Detail & Related papers  (2023-04-14T03:42:51Z)
• Deep HDR Hallucination for Inverse Tone Mapping [7.310237013012436]
  This work presents a GAN-based method that hallucinates missing information from badly exposed areas in LDR images. It provides good dynamic range expansion for well-exposed areas and plausible hallucinations for saturated and under-exposed areas.
  arXiv  Detail & Related papers  (2021-06-17T13:35:40Z)
• HDR-GAN: HDR Image Reconstruction from Multi-Exposed LDR Images with Large Motions [62.44802076971331]
  We propose a novel GAN-based model, HDR-GAN, for synthesizing HDR images from multi-exposed LDR images. By incorporating adversarial learning, our method is able to produce faithful information in the regions with missing content.
  arXiv  Detail & Related papers  (2020-07-03T11:42:35Z)
• Single-Image HDR Reconstruction by Learning to Reverse the Camera Pipeline [100.5353614588565]
  We propose to incorporate the domain knowledge of the LDR image formation pipeline into our model. We model the HDRto-LDR image formation pipeline as the (1) dynamic range clipping, (2) non-linear mapping from a camera response function, and (3) quantization. We demonstrate that the proposed method performs favorably against state-of-the-art single-image HDR reconstruction algorithms.
  arXiv  Detail & Related papers  (2020-04-02T17:59:04Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.