Coarse-to-fine Method for Vision-based Pedestrian Traffic Light Detection
Pedestrian traffic light detection is an important technique of the navigation system for the visually impaired during road crossing. In this paper, a three-stage coarse-to-fine method for pedestrian traffic light detection is proposed. The proposed method is mainly divided into two processes, the...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Stefan cel Mare University of Suceava
2020-02-01
|
Series: | Advances in Electrical and Computer Engineering |
Subjects: | |
Online Access: | http://dx.doi.org/10.4316/AECE.2020.01006 |
id |
doaj-b77f714e792a4bac9e085e5bc28a716f |
---|---|
record_format |
Article |
spelling |
doaj-b77f714e792a4bac9e085e5bc28a716f2020-11-25T01:11:13ZengStefan cel Mare University of SuceavaAdvances in Electrical and Computer Engineering1582-74451844-76002020-02-01201434810.4316/AECE.2020.01006Coarse-to-fine Method for Vision-based Pedestrian Traffic Light DetectionWU, X.-H.HU, R.BAO, Y.-Q.Pedestrian traffic light detection is an important technique of the navigation system for the visually impaired during road crossing. In this paper, a three-stage coarse-to-fine method for pedestrian traffic light detection is proposed. The proposed method is mainly divided into two processes, the training process and the detection process. In the training process, the Gaussian mixture model (GMM) is adopted to determine the parameters of the filter on stage I. The classifier on stage II is trained by a modified convolutional neural network (CNN) to capture features in each channel of the CIELAB color space. The classifier on stage III is trained by the adaptive boosting (AdaBoost) algorithm with Haar features. In the detection process, firstly the board filter is adopted to generate candidate regions of pedestrian traffic lights. Secondly, these candidate regions are detected in multiple scales by the CNN-based classifier with fixed size. Finally the AdaBoost-based classifier is adopted for refinement detection. Testing results verify the effectiveness of the proposed method.http://dx.doi.org/10.4316/AECE.2020.01006gaussian mixture modelmulti-layer neural networkboostingobject detectioncomputer vision |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
WU, X.-H. HU, R. BAO, Y.-Q. |
spellingShingle |
WU, X.-H. HU, R. BAO, Y.-Q. Coarse-to-fine Method for Vision-based Pedestrian Traffic Light Detection Advances in Electrical and Computer Engineering gaussian mixture model multi-layer neural network boosting object detection computer vision |
author_facet |
WU, X.-H. HU, R. BAO, Y.-Q. |
author_sort |
WU, X.-H. |
title |
Coarse-to-fine Method for Vision-based Pedestrian Traffic Light Detection |
title_short |
Coarse-to-fine Method for Vision-based Pedestrian Traffic Light Detection |
title_full |
Coarse-to-fine Method for Vision-based Pedestrian Traffic Light Detection |
title_fullStr |
Coarse-to-fine Method for Vision-based Pedestrian Traffic Light Detection |
title_full_unstemmed |
Coarse-to-fine Method for Vision-based Pedestrian Traffic Light Detection |
title_sort |
coarse-to-fine method for vision-based pedestrian traffic light detection |
publisher |
Stefan cel Mare University of Suceava |
series |
Advances in Electrical and Computer Engineering |
issn |
1582-7445 1844-7600 |
publishDate |
2020-02-01 |
description |
Pedestrian traffic light detection is an important technique of the navigation system for the visually impaired
during road crossing. In this paper, a three-stage coarse-to-fine method for pedestrian traffic light detection
is proposed. The proposed method is mainly divided into two processes, the training process and the detection
process. In the training process, the Gaussian mixture model (GMM) is adopted to determine the parameters of
the filter on stage I. The classifier on stage II is trained by a modified convolutional neural network (CNN)
to capture features in each channel of the CIELAB color space. The classifier on stage III is trained by the
adaptive boosting (AdaBoost) algorithm with Haar features. In the detection process, firstly the board filter
is adopted to generate candidate regions of pedestrian traffic lights. Secondly, these candidate regions are
detected in multiple scales by the CNN-based classifier with fixed size. Finally the AdaBoost-based classifier
is adopted for refinement detection. Testing results verify the effectiveness of the proposed method. |
topic |
gaussian mixture model multi-layer neural network boosting object detection computer vision |
url |
http://dx.doi.org/10.4316/AECE.2020.01006 |
work_keys_str_mv |
AT wuxh coarsetofinemethodforvisionbasedpedestriantrafficlightdetection AT hur coarsetofinemethodforvisionbasedpedestriantrafficlightdetection AT baoyq coarsetofinemethodforvisionbasedpedestriantrafficlightdetection |
_version_ |
1725172278413492224 |