Detecting pedestrians and vehicles in traffic scene based on boosted HOG features and SVM

Diqing Sun, Junzo Watada

Research output: Chapter in Book/Report/Conference proceedingConference contribution

14 Citations (Scopus)

Abstract

This paper presents a popular method called boosted hog features to detect pedestrians and vehicles in static images. We compared the differences and similarities of detecting pedestrians and vehicles, then we selected boosted hog features to get an satisfying result. In the part of detecting pedestrians, Histograms of Oriented Gradients (HOG) feature is applied as the basic feature due to its good performance in various kinds of background. On that basis, we create a new feature with boosting algorithm to obtain more accurate results. In the part of detecting vehicle, we select the shadow underneath vehicle as the feature, so we can utilize it to detect vehicles in daytime. The shadow is an important feature for vehicles in traffic scenes. The region under vehicle is usually darker than other objects or backgrounds and can be segmented by setting a threshold. Finally, experimental results show that the detection of pedestrians and vehicles using boosted hog feature and linear svm combines the advantages of hog feature and adaboost classifier, and can achieve better detection results than the detector using conditional HOG features. At the end, the paper shows its efficiency and effectiveness using to application in real situations.

Original languageEnglish
Title of host publicationWISP 2015 - IEEE International Symposium on Intelligent Signal Processing, Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Print)9781479972524
DOIs
Publication statusPublished - 2015 Jun 29
Externally publishedYes
Event9th IEEE International Symposium on Intelligent Signal Processing, WISP 2015 - Siena, Italy
Duration: 2015 May 152015 May 17

Other

Other9th IEEE International Symposium on Intelligent Signal Processing, WISP 2015
CountryItaly
CitySiena
Period15/5/1515/5/17

Fingerprint

Adaptive boosting
Classifiers
Detectors

ASJC Scopus subject areas

  • Artificial Intelligence
  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Sun, D., & Watada, J. (2015). Detecting pedestrians and vehicles in traffic scene based on boosted HOG features and SVM. In WISP 2015 - IEEE International Symposium on Intelligent Signal Processing, Proceedings [7139161] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/WISP.2015.7139161

Detecting pedestrians and vehicles in traffic scene based on boosted HOG features and SVM. / Sun, Diqing; Watada, Junzo.

WISP 2015 - IEEE International Symposium on Intelligent Signal Processing, Proceedings. Institute of Electrical and Electronics Engineers Inc., 2015. 7139161.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Sun, D & Watada, J 2015, Detecting pedestrians and vehicles in traffic scene based on boosted HOG features and SVM. in WISP 2015 - IEEE International Symposium on Intelligent Signal Processing, Proceedings., 7139161, Institute of Electrical and Electronics Engineers Inc., 9th IEEE International Symposium on Intelligent Signal Processing, WISP 2015, Siena, Italy, 15/5/15. https://doi.org/10.1109/WISP.2015.7139161
Sun D, Watada J. Detecting pedestrians and vehicles in traffic scene based on boosted HOG features and SVM. In WISP 2015 - IEEE International Symposium on Intelligent Signal Processing, Proceedings. Institute of Electrical and Electronics Engineers Inc. 2015. 7139161 https://doi.org/10.1109/WISP.2015.7139161
Sun, Diqing ; Watada, Junzo. / Detecting pedestrians and vehicles in traffic scene based on boosted HOG features and SVM. WISP 2015 - IEEE International Symposium on Intelligent Signal Processing, Proceedings. Institute of Electrical and Electronics Engineers Inc., 2015.
@inproceedings{7d0c4f6e64dd4f4da67cedc4173a9b9f,
title = "Detecting pedestrians and vehicles in traffic scene based on boosted HOG features and SVM",
abstract = "This paper presents a popular method called boosted hog features to detect pedestrians and vehicles in static images. We compared the differences and similarities of detecting pedestrians and vehicles, then we selected boosted hog features to get an satisfying result. In the part of detecting pedestrians, Histograms of Oriented Gradients (HOG) feature is applied as the basic feature due to its good performance in various kinds of background. On that basis, we create a new feature with boosting algorithm to obtain more accurate results. In the part of detecting vehicle, we select the shadow underneath vehicle as the feature, so we can utilize it to detect vehicles in daytime. The shadow is an important feature for vehicles in traffic scenes. The region under vehicle is usually darker than other objects or backgrounds and can be segmented by setting a threshold. Finally, experimental results show that the detection of pedestrians and vehicles using boosted hog feature and linear svm combines the advantages of hog feature and adaboost classifier, and can achieve better detection results than the detector using conditional HOG features. At the end, the paper shows its efficiency and effectiveness using to application in real situations.",
author = "Diqing Sun and Junzo Watada",
year = "2015",
month = "6",
day = "29",
doi = "10.1109/WISP.2015.7139161",
language = "English",
isbn = "9781479972524",
booktitle = "WISP 2015 - IEEE International Symposium on Intelligent Signal Processing, Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Detecting pedestrians and vehicles in traffic scene based on boosted HOG features and SVM

AU - Sun, Diqing

AU - Watada, Junzo

PY - 2015/6/29

Y1 - 2015/6/29

N2 - This paper presents a popular method called boosted hog features to detect pedestrians and vehicles in static images. We compared the differences and similarities of detecting pedestrians and vehicles, then we selected boosted hog features to get an satisfying result. In the part of detecting pedestrians, Histograms of Oriented Gradients (HOG) feature is applied as the basic feature due to its good performance in various kinds of background. On that basis, we create a new feature with boosting algorithm to obtain more accurate results. In the part of detecting vehicle, we select the shadow underneath vehicle as the feature, so we can utilize it to detect vehicles in daytime. The shadow is an important feature for vehicles in traffic scenes. The region under vehicle is usually darker than other objects or backgrounds and can be segmented by setting a threshold. Finally, experimental results show that the detection of pedestrians and vehicles using boosted hog feature and linear svm combines the advantages of hog feature and adaboost classifier, and can achieve better detection results than the detector using conditional HOG features. At the end, the paper shows its efficiency and effectiveness using to application in real situations.

AB - This paper presents a popular method called boosted hog features to detect pedestrians and vehicles in static images. We compared the differences and similarities of detecting pedestrians and vehicles, then we selected boosted hog features to get an satisfying result. In the part of detecting pedestrians, Histograms of Oriented Gradients (HOG) feature is applied as the basic feature due to its good performance in various kinds of background. On that basis, we create a new feature with boosting algorithm to obtain more accurate results. In the part of detecting vehicle, we select the shadow underneath vehicle as the feature, so we can utilize it to detect vehicles in daytime. The shadow is an important feature for vehicles in traffic scenes. The region under vehicle is usually darker than other objects or backgrounds and can be segmented by setting a threshold. Finally, experimental results show that the detection of pedestrians and vehicles using boosted hog feature and linear svm combines the advantages of hog feature and adaboost classifier, and can achieve better detection results than the detector using conditional HOG features. At the end, the paper shows its efficiency and effectiveness using to application in real situations.

UR - http://www.scopus.com/inward/record.url?scp=84943328867&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84943328867&partnerID=8YFLogxK

U2 - 10.1109/WISP.2015.7139161

DO - 10.1109/WISP.2015.7139161

M3 - Conference contribution

AN - SCOPUS:84943328867

SN - 9781479972524

BT - WISP 2015 - IEEE International Symposium on Intelligent Signal Processing, Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -