ETDNet: An Efficient Transformer Deraining Model

Rainy days usually degrade the visual effect of images and videos. At present, most deraining models for single images adopt gradual optimization or elimination to remove rain streaks, but actually with relatively low efficiency in real tasks. An efficient one-stage deraining model, Efficient Transf...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Qin Qin, Jingke Yan, Qin Wang, Xin Wang, Minyao Li, Yuqing Wang
Formato: article
Lenguaje:EN
Publicado: IEEE 2021
Materias:
Acceso en línea:https://doaj.org/article/511042f6fc5b44a98ef222d8f3bf542f
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:511042f6fc5b44a98ef222d8f3bf542f
record_format dspace
spelling oai:doaj.org-article:511042f6fc5b44a98ef222d8f3bf542f2021-11-19T00:07:02ZETDNet: An Efficient Transformer Deraining Model2169-353610.1109/ACCESS.2021.3108516https://doaj.org/article/511042f6fc5b44a98ef222d8f3bf542f2021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9524626/https://doaj.org/toc/2169-3536Rainy days usually degrade the visual effect of images and videos. At present, most deraining models for single images adopt gradual optimization or elimination to remove rain streaks, but actually with relatively low efficiency in real tasks. An efficient one-stage deraining model, Efficient Transformer Derain Network (ETDNet), is proposed to remove rain streaks in single images efficiently. A new Transformer architecture is designed to provide rich multiple scales and context information, making the model extract features in a coarse-to-fine way. Multiple expansion filters with different expansion rates are embedded to predict the suitable kernel for each pixel of the rainy image in a multi-scale way. A multi-scale Loss Function is introduced to restore features with high-fidelity and detail textures. Experiments on Rain100L, Rain100H, and SPA datasets show that the proposed ETDNet reaches the highest PSNR and SSIM values at the fastest speed compared with other models.Qin QinJingke YanQin WangXin WangMinyao LiYuqing WangIEEEarticleRain removalETDNettransformermulti-scaleloss functionElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 119881-119893 (2021)
institution DOAJ
collection DOAJ
language EN
topic Rain removal
ETDNet
transformer
multi-scale
loss function
Electrical engineering. Electronics. Nuclear engineering
TK1-9971
spellingShingle Rain removal
ETDNet
transformer
multi-scale
loss function
Electrical engineering. Electronics. Nuclear engineering
TK1-9971
Qin Qin
Jingke Yan
Qin Wang
Xin Wang
Minyao Li
Yuqing Wang
ETDNet: An Efficient Transformer Deraining Model
description Rainy days usually degrade the visual effect of images and videos. At present, most deraining models for single images adopt gradual optimization or elimination to remove rain streaks, but actually with relatively low efficiency in real tasks. An efficient one-stage deraining model, Efficient Transformer Derain Network (ETDNet), is proposed to remove rain streaks in single images efficiently. A new Transformer architecture is designed to provide rich multiple scales and context information, making the model extract features in a coarse-to-fine way. Multiple expansion filters with different expansion rates are embedded to predict the suitable kernel for each pixel of the rainy image in a multi-scale way. A multi-scale Loss Function is introduced to restore features with high-fidelity and detail textures. Experiments on Rain100L, Rain100H, and SPA datasets show that the proposed ETDNet reaches the highest PSNR and SSIM values at the fastest speed compared with other models.
format article
author Qin Qin
Jingke Yan
Qin Wang
Xin Wang
Minyao Li
Yuqing Wang
author_facet Qin Qin
Jingke Yan
Qin Wang
Xin Wang
Minyao Li
Yuqing Wang
author_sort Qin Qin
title ETDNet: An Efficient Transformer Deraining Model
title_short ETDNet: An Efficient Transformer Deraining Model
title_full ETDNet: An Efficient Transformer Deraining Model
title_fullStr ETDNet: An Efficient Transformer Deraining Model
title_full_unstemmed ETDNet: An Efficient Transformer Deraining Model
title_sort etdnet: an efficient transformer deraining model
publisher IEEE
publishDate 2021
url https://doaj.org/article/511042f6fc5b44a98ef222d8f3bf542f
work_keys_str_mv AT qinqin etdnetanefficienttransformerderainingmodel
AT jingkeyan etdnetanefficienttransformerderainingmodel
AT qinwang etdnetanefficienttransformerderainingmodel
AT xinwang etdnetanefficienttransformerderainingmodel
AT minyaoli etdnetanefficienttransformerderainingmodel
AT yuqingwang etdnetanefficienttransformerderainingmodel
_version_ 1718420608258146304