X2HDR: HDR Image Generation in a Perceptually Uniform Space

2026-02-04

AI Summary

Existing image generation models struggle to directly produce high dynamic range (HDR) images, primarily due to the scarcity of large-scale HDR datasets and the statistical discrepancies between sRGB and linear RGB domains. This work proposes mapping HDR images into perceptually uniform color spaces—such as PU21 or PQ—and leveraging a pre-trained low dynamic range (LDR) variational autoencoder (VAE) while applying only low-rank adaptation (LoRA) to the denoising network. This approach enables both text-to-HDR synthesis and single-image RAW-to-HDR reconstruction without full retraining. Notably, it demonstrates for the first time that perceptually uniform encoding effectively bridges the gap between LDR and HDR domains. By keeping the VAE frozen, the method unifies support for diverse HDR generation tasks and substantially improves perceptual fidelity, text alignment, and effective dynamic range over existing techniques.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Natural Language Processing: Generation

Application Category

• Search and Retrieval-Augmented AI: Web search models and ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

High-dynamic-range (HDR) formats and displays are becoming increasingly prevalent, yet state-of-the-art image generators (e.g., Stable Diffusion and FLUX) typically remain limited to low-dynamic-range (LDR) output due to the lack of large-scale HDR training data. In this work, we show that existing pretrained diffusion models can be easily adapted to HDR generation without retraining from scratch. A key challenge is that HDR images are natively represented in linear RGB, whose intensity and color statistics differ substantially from those of sRGB-encoded LDR images. This gap, however, can be effectively bridged by converting HDR inputs into perceptually uniform encodings (e.g., using PU21 or PQ). Empirically, we find that LDR-pretrained variational autoencoders (VAEs) reconstruct PU21-encoded HDR inputs with fidelity comparable to LDR data, whereas linear RGB inputs cause severe degradations. Motivated by this finding, we describe an efficient adaptation strategy that freezes the VAE and finetunes only the denoiser via low-rank adaptation in a perceptually uniform space. This results in a unified computational method that supports both text-to-HDR synthesis and single-image RAW-to-HDR reconstruction. Experiments demonstrate that our perceptually encoded adaptation consistently improves perceptual fidelity, text-image alignment, and effective dynamic range, relative to previous techniques.

Problem

• HDR image generation
• perceptually uniform space
• diffusion models
• dynamic range
• image synthesis

Innovation

• perceptually uniform encoding
• HDR image generation
• low-rank adaptation
• diffusion models
• VAE reconstruction