Deep Neural Networks for HDR imaging

TL;DR

This paper introduces neural network-based methods for generating HDR images from multiple exposures and for optimizing tone mapping operators, demonstrating improved performance on quality metrics.

Contribution

It presents novel CNN approaches for HDR image synthesis and tone mapping optimization, advancing the state-of-the-art in HDR imaging techniques.

Findings

CNN methods outperform existing techniques on TMQI metric

Networks effectively generate HDR maps from LDR images

Performance varies across different scene conditions

Abstract

We propose novel methods of solving two tasks using Convolutional Neural Networks, firstly the task of generating HDR map of a static scene using differently exposed LDR images of the scene captured using conventional cameras and secondly the task of finding an optimal tone mapping operator that would give a better score on the TMQI metric compared to the existing methods. We quantitatively show the performance of our networks and illustrate the cases where our networks performs good as well as bad.