Issues with common assumptions in HDR imaging from multiple exposures

TitleIssues with common assumptions in HDR imaging from multiple exposures
Publication TypeJournal Article
Year of Publication2018
AuthorsRodríguez RGil, Vazquez-Corral J, Bertalmío M
Journal*Submitted* to Transactions on Pattern Analysis and Machine Intelligence (TPAMI)
Abstract

Multiple-exposure approaches for HDR image generation share a set of building assumptions: that the full dynamic range of the scene can be recovered, that color channels are independent, and that the camera response function remains constant while changing the exposure. The contributions of this paper are, on one hand, to highlight how these assumptions do not hold in general, and then to propose a method to improve multiple-exposure combination, compensating for the violations of the latter assumptions and allowing us to obtain HDR results, both for static and dynamic scenes, that are shown to be more accurate to the ground truth HDR image (derived directly from RAW pictures) than those obtained with state-of-the-art multiple-exposure combination methods, according to several image metrics. And, very importantly, the tone-mapped versions of the HDR pictures obtained with our method show no hue shifts, color artifacts or contrast problems, which are issues that commonly crop up with other multiple-exposure approaches.