DRHDR: A Dual Branch Residual Network for Multi-Bracket High Dynamic Range Imaging
Abstract
We introduce DRHDR, a Dual branch Residual Convolutional Neural Network for Multi-Bracket HDR Imaging. To address the challenges of fusing multiple brackets from dynamic scenes, we propose an efficient dual branch network that operates on two different resolutions. The full resolution branch uses a Deformable Convolutional Block to align features and retain high-frequency details. A low resolution branch with a Spatial Attention Block aims to attend wanted areas from the non-reference brackets, and suppress displaced features that could incur on ghosting artifacts. By using a dual branch approach we are able to achieve high quality results while constraining the computational resources required to estimate the HDR results.
Cite
Text
Marín-Vega et al. "DRHDR: A Dual Branch Residual Network for Multi-Bracket High Dynamic Range Imaging." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2022. doi:10.1109/CVPRW56347.2022.00100Markdown
[Marín-Vega et al. "DRHDR: A Dual Branch Residual Network for Multi-Bracket High Dynamic Range Imaging." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2022.](https://mlanthology.org/cvprw/2022/marinvega2022cvprw-drhdr/) doi:10.1109/CVPRW56347.2022.00100BibTeX
@inproceedings{marinvega2022cvprw-drhdr,
title = {{DRHDR: A Dual Branch Residual Network for Multi-Bracket High Dynamic Range Imaging}},
author = {Marín-Vega, Juan and Sloth, Michael and Schneider-Kamp, Peter and Röttger, Richard},
booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},
year = {2022},
pages = {843-851},
doi = {10.1109/CVPRW56347.2022.00100},
url = {https://mlanthology.org/cvprw/2022/marinvega2022cvprw-drhdr/}
}