Optimal HDR Reconstruction with Linear Digital Cameras
Abstract
Given a multi-exposure sequence of a scene, our aim is to recover the absolute \nirradiance falling onto a linear camera sensor. The established approach is to \nperform a weighted average of the scaled input exposures. However, there is no \nclear consensus on the appropriate weighting to use. We propose a weighting \nfunction that produces statistically optimal estimates under the assumption of \ncompound- Gaussian noise. Our weighting is based on a calibrated camera model \nthat accounts for all noise sources. This model also allows us to \nsimultaneously estimate the irradiance and its uncertainty. We evaluate our \nmethod on simulated and real world photographs, and show that we consistently \nimprove the signal-to-noise ratio over previous approaches. Finally, we show \nthe effectiveness of our model for optimal exposure sequence selection and HDR \nimage denoising.
Cite
Text
Granados et al. "Optimal HDR Reconstruction with Linear Digital Cameras." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2010. doi:10.1109/CVPR.2010.5540208Markdown
[Granados et al. "Optimal HDR Reconstruction with Linear Digital Cameras." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2010.](https://mlanthology.org/cvpr/2010/granados2010cvpr-optimal/) doi:10.1109/CVPR.2010.5540208BibTeX
@inproceedings{granados2010cvpr-optimal,
title = {{Optimal HDR Reconstruction with Linear Digital Cameras}},
author = {Granados, Miguel and Ajdin, Boris and Wand, Michael and Theobalt, Christian and Seidel, Hans-Peter and Lensch, Hendrik P. A.},
booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
year = {2010},
pages = {215-222},
doi = {10.1109/CVPR.2010.5540208},
url = {https://mlanthology.org/cvpr/2010/granados2010cvpr-optimal/}
}