HDR Reconstruction Boosting with Training-Free and Exposure-Consistent Diffusion

TL;DR

This paper introduces a training-free diffusion-based method that enhances HDR reconstruction from multiple LDR images, especially in over-exposed regions, by improving detail recovery and luminance consistency.

Contribution

It presents a novel, training-free approach combining diffusion models with existing HDR methods to improve over-exposed region reconstruction and multi-exposure consistency.

Findings

Significant perceptual quality improvements on standard HDR datasets.

Effective recovery of details in over-exposed regions.

Seamless integration with existing HDR pipelines.

Abstract

Single LDR to HDR reconstruction remains challenging for over-exposed regions where traditional methods often fail due to complete information loss. We present a training-free approach that enhances existing indirect and direct HDR reconstruction methods through diffusion-based inpainting. Our method combines text-guided diffusion models with SDEdit refinement to generate plausible content in over-exposed areas while maintaining consistency across multi-exposure LDR images. Unlike previous approaches requiring extensive training, our method seamlessly integrates with existing HDR reconstruction techniques through an iterative compensation mechanism that ensures luminance coherence across multiple exposures. We demonstrate significant improvements in both perceptual quality and quantitative metrics on standard HDR datasets and in-the-wild captures. Results show that our method effectively recovers natural details in challenging scenarios while preserving the advantages of existing HDR reconstruction pipelines. Project page: https://github.com/EusdenLin/HDR-Reconstruction-Boosting