The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
En tant que point névralgique de la recherche et difficulté dans le domaine de la vision par ordinateur, la détection des piétons a été largement utilisée dans la conduite intelligente et la surveillance du trafic. La méthode de détection populaire à l'heure actuelle utilise un réseau de proposition de région (RPN) pour générer des régions candidates, puis classe les régions. Mais le RPN produit de nombreuses zones candidates erronées, ce qui entraîne une augmentation des propositions régionales de faux positifs. Cette lettre utilise un réseau d’attention résiduelle amélioré pour capturer la carte d’attention visuelle des images, puis normalisée pour obtenir la carte des scores d’attention. La carte des scores d'attention est utilisée pour guider le réseau RPN afin de générer des régions candidates plus précises contenant des objets cibles potentiels. Les propositions de région, les scores de confiance et les fonctionnalités générés par le RPN sont utilisés pour former un classificateur forestier boosté en cascade afin d'obtenir les résultats finaux. Les résultats expérimentaux montrent que notre approche proposée permet d'obtenir des résultats très compétitifs sur les ensembles de données Caltech et ETH.
Rui SUN
Hefei University of Technology
Huihui WANG
Hefei University of Technology
Jun ZHANG
Hefei University of Technology
Xudong ZHANG
Hefei University of Technology
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.
Copier
Rui SUN, Huihui WANG, Jun ZHANG, Xudong ZHANG, "Attention-Guided Region Proposal Network for Pedestrian Detection" in IEICE TRANSACTIONS on Information,
vol. E102-D, no. 10, pp. 2072-2076, October 2019, doi: 10.1587/transinf.2019EDL8027.
Abstract: As a research hotspot and difficulty in the field of computer vision, pedestrian detection has been widely used in intelligent driving and traffic monitoring. The popular detection method at present uses region proposal network (RPN) to generate candidate regions, and then classifies the regions. But the RPN produces many erroneous candidate areas, causing region proposals for false positives to increase. This letter uses improved residual attention network to capture the visual attention map of images, then normalized to get the attention score map. The attention score map is used to guide the RPN network to generate more precise candidate regions containing potential target objects. The region proposals, confidence scores, and features generated by the RPN are used to train a cascaded boosted forest classifier to obtain the final results. The experimental results show that our proposed approach achieves highly competitive results on the Caltech and ETH datasets.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2019EDL8027/_p
Copier
@ARTICLE{e102-d_10_2072,
author={Rui SUN, Huihui WANG, Jun ZHANG, Xudong ZHANG, },
journal={IEICE TRANSACTIONS on Information},
title={Attention-Guided Region Proposal Network for Pedestrian Detection},
year={2019},
volume={E102-D},
number={10},
pages={2072-2076},
abstract={As a research hotspot and difficulty in the field of computer vision, pedestrian detection has been widely used in intelligent driving and traffic monitoring. The popular detection method at present uses region proposal network (RPN) to generate candidate regions, and then classifies the regions. But the RPN produces many erroneous candidate areas, causing region proposals for false positives to increase. This letter uses improved residual attention network to capture the visual attention map of images, then normalized to get the attention score map. The attention score map is used to guide the RPN network to generate more precise candidate regions containing potential target objects. The region proposals, confidence scores, and features generated by the RPN are used to train a cascaded boosted forest classifier to obtain the final results. The experimental results show that our proposed approach achieves highly competitive results on the Caltech and ETH datasets.},
keywords={},
doi={10.1587/transinf.2019EDL8027},
ISSN={1745-1361},
month={October},}
Copier
TY - JOUR
TI - Attention-Guided Region Proposal Network for Pedestrian Detection
T2 - IEICE TRANSACTIONS on Information
SP - 2072
EP - 2076
AU - Rui SUN
AU - Huihui WANG
AU - Jun ZHANG
AU - Xudong ZHANG
PY - 2019
DO - 10.1587/transinf.2019EDL8027
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E102-D
IS - 10
JA - IEICE TRANSACTIONS on Information
Y1 - October 2019
AB - As a research hotspot and difficulty in the field of computer vision, pedestrian detection has been widely used in intelligent driving and traffic monitoring. The popular detection method at present uses region proposal network (RPN) to generate candidate regions, and then classifies the regions. But the RPN produces many erroneous candidate areas, causing region proposals for false positives to increase. This letter uses improved residual attention network to capture the visual attention map of images, then normalized to get the attention score map. The attention score map is used to guide the RPN network to generate more precise candidate regions containing potential target objects. The region proposals, confidence scores, and features generated by the RPN are used to train a cascaded boosted forest classifier to obtain the final results. The experimental results show that our proposed approach achieves highly competitive results on the Caltech and ETH datasets.
ER -