X. Luo, J. Peng, K. Xian, Z. Wu, and Z. Cao. Computer Vision -- ECCV 2020 Workshops, page 245--261. Cham, Springer International Publishing, (2020)
Abstract
In this paper, we study realistic bokeh rendering from a single all-in-focus image. Existing computational bokeh rendering methods generate bokeh effects by adding a simple flat background blur. As a result, the rendering results are different from the real bokeh on DSLR cameras. To address this issue, we propose a multi-stage network to learn shallow depth-of-field from a single bokeh-free image. In particular, our network consists of four modules: defocus estimation, radiance, rendering, and upsampling. The four modules are trained on different sizes to learn global features as well as local details around the boundaries of in-focus objects. Experimental results show that our approach is capable of rendering a pleasing distinctive bokeh effect in complex scenes.
Description
Bokeh Rendering from Defocus Estimation | SpringerLink
%0 Conference Paper
%1 LuoDefucsEstimation
%A Luo, Xianrui
%A Peng, Juewen
%A Xian, Ke
%A Wu, Zijin
%A Cao, Zhiguo
%B Computer Vision -- ECCV 2020 Workshops
%C Cham
%D 2020
%E Bartoli, Adrien
%E Fusiello, Andrea
%I Springer International Publishing
%K cv_bokehseminar
%P 245--261
%T Bokeh Rendering from Defocus Estimation
%X In this paper, we study realistic bokeh rendering from a single all-in-focus image. Existing computational bokeh rendering methods generate bokeh effects by adding a simple flat background blur. As a result, the rendering results are different from the real bokeh on DSLR cameras. To address this issue, we propose a multi-stage network to learn shallow depth-of-field from a single bokeh-free image. In particular, our network consists of four modules: defocus estimation, radiance, rendering, and upsampling. The four modules are trained on different sizes to learn global features as well as local details around the boundaries of in-focus objects. Experimental results show that our approach is capable of rendering a pleasing distinctive bokeh effect in complex scenes.
%@ 978-3-030-67070-2
@inproceedings{LuoDefucsEstimation,
abstract = {In this paper, we study realistic bokeh rendering from a single all-in-focus image. Existing computational bokeh rendering methods generate bokeh effects by adding a simple flat background blur. As a result, the rendering results are different from the real bokeh on DSLR cameras. To address this issue, we propose a multi-stage network to learn shallow depth-of-field from a single bokeh-free image. In particular, our network consists of four modules: defocus estimation, radiance, rendering, and upsampling. The four modules are trained on different sizes to learn global features as well as local details around the boundaries of in-focus objects. Experimental results show that our approach is capable of rendering a pleasing distinctive bokeh effect in complex scenes.},
added-at = {2022-12-18T09:40:09.000+0100},
address = {Cham},
author = {Luo, Xianrui and Peng, Juewen and Xian, Ke and Wu, Zijin and Cao, Zhiguo},
biburl = {https://www.bibsonomy.org/bibtex/2d2c28f0fa370d9efa2b855d2f22c0410/t_seizinger},
booktitle = {Computer Vision -- ECCV 2020 Workshops},
description = {Bokeh Rendering from Defocus Estimation | SpringerLink},
editor = {Bartoli, Adrien and Fusiello, Andrea},
interhash = {938e56de1576f0fd0c003935f2156388},
intrahash = {d2c28f0fa370d9efa2b855d2f22c0410},
isbn = {978-3-030-67070-2},
keywords = {cv_bokehseminar},
pages = {245--261},
publisher = {Springer International Publishing},
timestamp = {2022-12-18T09:40:09.000+0100},
title = {Bokeh Rendering from Defocus Estimation},
year = 2020
}