@inproceedings{13e028fed801423388b8e47ba8de8b64,
title = "Non-local exposure fusion",
abstract = "Exposure fusion is an efficient method to obtain a well exposed and detailed image from a scene with high dynamic range. However, this method fails when there is camera shake and/or object motions. In this work, we tackle this issue by replacing the pixel-based fusion by a fusion between pixels having similar neighborhood (patches) in images with different exposure settings. In order to achieve this, we compare patches in the luminance domain. We show through several experiments that this procedure yield comparable or better results than the state of the art, at a reasonable computing time.",
keywords = "Exposure fusion, High dynamic range, Image fusion, Non-local methods, Patches",
author = "Cristian Ocampo-Blandon and Yann Gousseau",
note = "Publisher Copyright: {\textcopyright} Springer International Publishing AG 2017.; 21st Iberoamerican Congress on Pattern Recognition, CIARP 2016 ; Conference date: 08-11-2016 Through 11-11-2016",
year = "2017",
month = jan,
day = "1",
doi = "10.1007/978-3-319-52277-7\_59",
language = "English",
isbn = "9783319522760",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "484--492",
editor = "Cesar Beltran-Castanon and Fazel Famili and Ingela Nystrom",
booktitle = "Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications - 21st Iberoamerican Congress, CIARP 2016, Proceedings",
}