AI RESEARCH
LatentHDR: Decoupling Exposure from Diffusion via Conditional Latent-to-Latent Mapping for Text/Image-to-Panoramic HDR
arXiv CS.LG
•
ArXi:2605.11115v1 Announce Type: cross High Dynamic Range (HDR) generation remains challenging for generative models, which are largely limited to low dynamic range outputs. Recent diffusionbased approaches approximate HDR by generating multiple exposure-conditioned samples, incurring high computational cost and structural inconsistencies across exposures. We propose LatentHDR, a framework that decouples scene generation from exposure modeling in latent space.