In this paper, we propose an end-to-end two-branch feature attention network. The network is mainly used for single image dehazing. The network consists of two branches, we call it CAA-Net: 1) A U-NET network composed of different-level feature fusion based on attention (FEPA) structure and residual dense block (RDB). In order to make full use of all the hierarchical features of the image, we use RDB. RDB contains dense connected layers and local feature fusion with local residual learning. We also propose a structure which called FEPA.FEPA structure could retain the information of shallow layer and transfer it to the deep layer. FEPA is composed of serveral feature attention modules (FPA). FPA combines local residual learning with channel attention mechanism and pixel attention mechanism, and could extract features from different channels and image pixels. 2) A network composed of several different levels of FEPA structures. The network could make feature weights learn from FPA adaptively, and give more weight to important features. The final output result of CAA-Net is the combination of all branch prediction results. Experimental results show that the CAA-Net proposed by us surpasses the most advanced algorithms before for single image dehazing.
Gang JIN
Tianjin University
Jingsheng ZHAI
Tianjin University
Jianguo WEI
Tianjin University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Gang JIN, Jingsheng ZHAI, Jianguo WEI, "CAA-Net: End-to-End Two-Branch Feature Attention Network for Single Image Dehazing" in IEICE TRANSACTIONS on Fundamentals,
vol. E106-A, no. 1, pp. 1-10, January 2023, doi: 10.1587/transfun.2022EAP1019.
Abstract: In this paper, we propose an end-to-end two-branch feature attention network. The network is mainly used for single image dehazing. The network consists of two branches, we call it CAA-Net: 1) A U-NET network composed of different-level feature fusion based on attention (FEPA) structure and residual dense block (RDB). In order to make full use of all the hierarchical features of the image, we use RDB. RDB contains dense connected layers and local feature fusion with local residual learning. We also propose a structure which called FEPA.FEPA structure could retain the information of shallow layer and transfer it to the deep layer. FEPA is composed of serveral feature attention modules (FPA). FPA combines local residual learning with channel attention mechanism and pixel attention mechanism, and could extract features from different channels and image pixels. 2) A network composed of several different levels of FEPA structures. The network could make feature weights learn from FPA adaptively, and give more weight to important features. The final output result of CAA-Net is the combination of all branch prediction results. Experimental results show that the CAA-Net proposed by us surpasses the most advanced algorithms before for single image dehazing.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.2022EAP1019/_p
Copy
@ARTICLE{e106-a_1_1,
author={Gang JIN, Jingsheng ZHAI, Jianguo WEI, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={CAA-Net: End-to-End Two-Branch Feature Attention Network for Single Image Dehazing},
year={2023},
volume={E106-A},
number={1},
pages={1-10},
abstract={In this paper, we propose an end-to-end two-branch feature attention network. The network is mainly used for single image dehazing. The network consists of two branches, we call it CAA-Net: 1) A U-NET network composed of different-level feature fusion based on attention (FEPA) structure and residual dense block (RDB). In order to make full use of all the hierarchical features of the image, we use RDB. RDB contains dense connected layers and local feature fusion with local residual learning. We also propose a structure which called FEPA.FEPA structure could retain the information of shallow layer and transfer it to the deep layer. FEPA is composed of serveral feature attention modules (FPA). FPA combines local residual learning with channel attention mechanism and pixel attention mechanism, and could extract features from different channels and image pixels. 2) A network composed of several different levels of FEPA structures. The network could make feature weights learn from FPA adaptively, and give more weight to important features. The final output result of CAA-Net is the combination of all branch prediction results. Experimental results show that the CAA-Net proposed by us surpasses the most advanced algorithms before for single image dehazing.},
keywords={},
doi={10.1587/transfun.2022EAP1019},
ISSN={1745-1337},
month={January},}
Copy
TY - JOUR
TI - CAA-Net: End-to-End Two-Branch Feature Attention Network for Single Image Dehazing
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 1
EP - 10
AU - Gang JIN
AU - Jingsheng ZHAI
AU - Jianguo WEI
PY - 2023
DO - 10.1587/transfun.2022EAP1019
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E106-A
IS - 1
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - January 2023
AB - In this paper, we propose an end-to-end two-branch feature attention network. The network is mainly used for single image dehazing. The network consists of two branches, we call it CAA-Net: 1) A U-NET network composed of different-level feature fusion based on attention (FEPA) structure and residual dense block (RDB). In order to make full use of all the hierarchical features of the image, we use RDB. RDB contains dense connected layers and local feature fusion with local residual learning. We also propose a structure which called FEPA.FEPA structure could retain the information of shallow layer and transfer it to the deep layer. FEPA is composed of serveral feature attention modules (FPA). FPA combines local residual learning with channel attention mechanism and pixel attention mechanism, and could extract features from different channels and image pixels. 2) A network composed of several different levels of FEPA structures. The network could make feature weights learn from FPA adaptively, and give more weight to important features. The final output result of CAA-Net is the combination of all branch prediction results. Experimental results show that the CAA-Net proposed by us surpasses the most advanced algorithms before for single image dehazing.
ER -