FastDerainNet: A Deep Learning Algorithm for Single Image Deraining

Existing neural network-based methods for de-raining single images exhibit dissatisfactory results owing to the inefficient propagation of features when objects with sizes and shapes similar to those of rain streaks are present in images. Furthermore, existing methods do not consider that the abunda...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Xiuwen Wang, Zhiwei Li, Hongtao Shan, Zhiyuan Tian, Yuanhong Ren, Wuneng Zhou
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2020
Materias:	Deep residual learning rain streak removal FastDerainNet share-source residual module Electrical engineering. Electronics. Nuclear engineering TK1-9971
Acceso en línea:	https://doaj.org/article/24873bd045b2484eb0e879414213feba
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:24873bd045b2484eb0e879414213feba
record_format	dspace
spelling	oai:doaj.org-article:24873bd045b2484eb0e879414213feba2021-11-19T00:03:51ZFastDerainNet: A Deep Learning Algorithm for Single Image Deraining2169-353610.1109/ACCESS.2020.3008324https://doaj.org/article/24873bd045b2484eb0e879414213feba2020-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9139246/https://doaj.org/toc/2169-3536Existing neural network-based methods for de-raining single images exhibit dissatisfactory results owing to the inefficient propagation of features when objects with sizes and shapes similar to those of rain streaks are present in images. Furthermore, existing methods do not consider that the abundant information included in rain streaked images could interfere with the training process. To overcome these limitations, in this paper, we propose a deep residual learning algorithm called FastDerainNet for removing rain streaks from single images. We design a deep convolutional neural network architecture, based on a deep residual network called the share-source residual module (SSRM), by substituting the origins of all shortcut connections for one point. To further improve the de-raining performance, we adopt the SSRM as the parameter layers in FastDerainNet and use image decomposition to modify the loss function. Finally, we train FastDerainNet on a synthetic dataset. By learning the residual mapping between rainy and clean image detail layers, it is able to reduce the mapping range and simplify the training process. Experiments on both synthetic and real-world images demonstrate that the proposed method achieves increased performance with regard to de-raining, in addition to preserving original details, in comparison with other state-of-the-art methods.Xiuwen WangZhiwei LiHongtao ShanZhiyuan TianYuanhong RenWuneng ZhouIEEEarticleDeep residual learningrain streak removalFastDerainNetshare-source residual moduleElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 8, Pp 127622-127630 (2020)
institution	DOAJ
collection	DOAJ
language	EN
topic	Deep residual learning rain streak removal FastDerainNet share-source residual module Electrical engineering. Electronics. Nuclear engineering TK1-9971
spellingShingle	Deep residual learning rain streak removal FastDerainNet share-source residual module Electrical engineering. Electronics. Nuclear engineering TK1-9971 Xiuwen Wang Zhiwei Li Hongtao Shan Zhiyuan Tian Yuanhong Ren Wuneng Zhou FastDerainNet: A Deep Learning Algorithm for Single Image Deraining
description	Existing neural network-based methods for de-raining single images exhibit dissatisfactory results owing to the inefficient propagation of features when objects with sizes and shapes similar to those of rain streaks are present in images. Furthermore, existing methods do not consider that the abundant information included in rain streaked images could interfere with the training process. To overcome these limitations, in this paper, we propose a deep residual learning algorithm called FastDerainNet for removing rain streaks from single images. We design a deep convolutional neural network architecture, based on a deep residual network called the share-source residual module (SSRM), by substituting the origins of all shortcut connections for one point. To further improve the de-raining performance, we adopt the SSRM as the parameter layers in FastDerainNet and use image decomposition to modify the loss function. Finally, we train FastDerainNet on a synthetic dataset. By learning the residual mapping between rainy and clean image detail layers, it is able to reduce the mapping range and simplify the training process. Experiments on both synthetic and real-world images demonstrate that the proposed method achieves increased performance with regard to de-raining, in addition to preserving original details, in comparison with other state-of-the-art methods.
format	article
author	Xiuwen Wang Zhiwei Li Hongtao Shan Zhiyuan Tian Yuanhong Ren Wuneng Zhou
author_facet	Xiuwen Wang Zhiwei Li Hongtao Shan Zhiyuan Tian Yuanhong Ren Wuneng Zhou
author_sort	Xiuwen Wang
title	FastDerainNet: A Deep Learning Algorithm for Single Image Deraining
title_short	FastDerainNet: A Deep Learning Algorithm for Single Image Deraining
title_full	FastDerainNet: A Deep Learning Algorithm for Single Image Deraining
title_fullStr	FastDerainNet: A Deep Learning Algorithm for Single Image Deraining
title_full_unstemmed	FastDerainNet: A Deep Learning Algorithm for Single Image Deraining
title_sort	fastderainnet: a deep learning algorithm for single image deraining
publisher	IEEE
publishDate	2020
url	https://doaj.org/article/24873bd045b2484eb0e879414213feba
work_keys_str_mv	AT xiuwenwang fastderainnetadeeplearningalgorithmforsingleimagederaining AT zhiweili fastderainnetadeeplearningalgorithmforsingleimagederaining AT hongtaoshan fastderainnetadeeplearningalgorithmforsingleimagederaining AT zhiyuantian fastderainnetadeeplearningalgorithmforsingleimagederaining AT yuanhongren fastderainnetadeeplearningalgorithmforsingleimagederaining AT wunengzhou fastderainnetadeeplearningalgorithmforsingleimagederaining
_version_	1718420657208819712

FastDerainNet: A Deep Learning Algorithm for Single Image Deraining

Ejemplares similares