Feature Enhancement Based on CycleGAN for Nighttime Vehicle Detection
Existing night vehicle detection methods mainly detect vehicles by detecting headlights or taillights. However, these features are adversely affected by the complex road lighting environment. In this paper, a cascade detection network framework FteGanOd is proposed with a feature translate-enhanceme...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2021-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/9303452/ |
id |
doaj-708b033335ad452da4fd3ab354362120 |
---|---|
record_format |
Article |
spelling |
doaj-708b033335ad452da4fd3ab3543621202021-03-30T14:56:49ZengIEEEIEEE Access2169-35362021-01-01984985910.1109/ACCESS.2020.30464989303452Feature Enhancement Based on CycleGAN for Nighttime Vehicle DetectionXiaotao Shao0https://orcid.org/0000-0003-0758-518XCaike Wei1https://orcid.org/0000-0002-3471-0014Yan Shen2https://orcid.org/0000-0001-9287-1206Zhongli Wang3https://orcid.org/0000-0002-3236-8219School of Electronic and Information Engineering, Beijing Jiaotong University, Beijing, ChinaSchool of Electronic and Information Engineering, Beijing Jiaotong University, Beijing, ChinaSchool of Electronic and Information Engineering, Beijing Jiaotong University, Beijing, ChinaSchool of Electronic and Information Engineering, Beijing Jiaotong University, Beijing, ChinaExisting night vehicle detection methods mainly detect vehicles by detecting headlights or taillights. However, these features are adversely affected by the complex road lighting environment. In this paper, a cascade detection network framework FteGanOd is proposed with a feature translate-enhancement (FTE) module and the object detection (OD) module. First, the FTE module is built based on CycleGAN and multi-scale feature fusion is proposed to enhance the detection of vehicle features at night. The features of night and day are combined by fusing different convolutional layers to produce enhanced feature (EF) maps. Second, the OD module, based on the existing object detection network, is improved by cascading with the FTE module to detect vehicles on the EF maps. The proposed FteGanOd method recognizes vehicles at night with greater accuracy by improving the contrast between vehicles and the background and by suppressing interference from ambient light. The proposed FteGanOd is validated on the Berkeley Deep Drive (BDD) dataset and our private dataset. The experimental results show that our proposed method can effectively enhance vehicle features and improve the accuracy of vehicle detection at night.https://ieeexplore.ieee.org/document/9303452/Convolutional neural network (CNN)nighttime vehicle detectionfeature enhancementgenerative adversarial network (GAN)detection network |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Xiaotao Shao Caike Wei Yan Shen Zhongli Wang |
spellingShingle |
Xiaotao Shao Caike Wei Yan Shen Zhongli Wang Feature Enhancement Based on CycleGAN for Nighttime Vehicle Detection IEEE Access Convolutional neural network (CNN) nighttime vehicle detection feature enhancement generative adversarial network (GAN) detection network |
author_facet |
Xiaotao Shao Caike Wei Yan Shen Zhongli Wang |
author_sort |
Xiaotao Shao |
title |
Feature Enhancement Based on CycleGAN for Nighttime Vehicle Detection |
title_short |
Feature Enhancement Based on CycleGAN for Nighttime Vehicle Detection |
title_full |
Feature Enhancement Based on CycleGAN for Nighttime Vehicle Detection |
title_fullStr |
Feature Enhancement Based on CycleGAN for Nighttime Vehicle Detection |
title_full_unstemmed |
Feature Enhancement Based on CycleGAN for Nighttime Vehicle Detection |
title_sort |
feature enhancement based on cyclegan for nighttime vehicle detection |
publisher |
IEEE |
series |
IEEE Access |
issn |
2169-3536 |
publishDate |
2021-01-01 |
description |
Existing night vehicle detection methods mainly detect vehicles by detecting headlights or taillights. However, these features are adversely affected by the complex road lighting environment. In this paper, a cascade detection network framework FteGanOd is proposed with a feature translate-enhancement (FTE) module and the object detection (OD) module. First, the FTE module is built based on CycleGAN and multi-scale feature fusion is proposed to enhance the detection of vehicle features at night. The features of night and day are combined by fusing different convolutional layers to produce enhanced feature (EF) maps. Second, the OD module, based on the existing object detection network, is improved by cascading with the FTE module to detect vehicles on the EF maps. The proposed FteGanOd method recognizes vehicles at night with greater accuracy by improving the contrast between vehicles and the background and by suppressing interference from ambient light. The proposed FteGanOd is validated on the Berkeley Deep Drive (BDD) dataset and our private dataset. The experimental results show that our proposed method can effectively enhance vehicle features and improve the accuracy of vehicle detection at night. |
topic |
Convolutional neural network (CNN) nighttime vehicle detection feature enhancement generative adversarial network (GAN) detection network |
url |
https://ieeexplore.ieee.org/document/9303452/ |
work_keys_str_mv |
AT xiaotaoshao featureenhancementbasedoncycleganfornighttimevehicledetection AT caikewei featureenhancementbasedoncycleganfornighttimevehicledetection AT yanshen featureenhancementbasedoncycleganfornighttimevehicledetection AT zhongliwang featureenhancementbasedoncycleganfornighttimevehicledetection |
_version_ |
1724180344239816704 |