Pedestrian detection system using cascaded boosting with invariance of oriented gradients

Chi Chen Raxle Wang, Jin Yi Wu, James Jenn-Jier Lien

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

This study presents a novel learning-based pedestrian detection system capable of automatically detecting individuals of different sizes and orientations against a wide variety of backgrounds, including crowds, even when the individual is partially occluded. To render the detection performance robust toward the effects of geometric and rotational variations in the original image, the feature extraction process is performed using both rectangular- and circular-type blocks of various sizes and aspect ratios. The extracted blocks are rotated in accordance with their dominant orientation(s) such that all the blocks extracted from the input images are rotationally invariant. The pixels within the cells in each block are then voted into rectangular- and circular-type 9-bin histograms of oriented gradients (HOGs) in accordance with their gradient magnitudes and corresponding multivariate Gaussian-weighted windows. Finally, four cell-based histograms are concatenated using a tri-linear interpolation technique to form one 36-dimensional normalized HOG feature vector for each block. The experimental results show that the use of the Gaussian-weighted window approach and tri-linear interpolation technique in constructing the HOG feature vectors improves the detection performance from 91% to 94.5%. In the proposed scheme, the detection process is performed using a cascaded detector structure in which the weak classifiers and corresponding weights of each stage are established using the AdaBoost self-learning algorithm. The experimental results reveal that the cascaded structure not only provides a better detection performance than many of the schemes presented in the literature, but also achieves a significant reduction in the computational time required to classify each input image.

Original languageEnglish
Pages (from-to)801-823
Number of pages23
JournalInternational Journal of Pattern Recognition and Artificial Intelligence
Volume23
Issue number4
DOIs
Publication statusPublished - 2009 Jun 1

Fingerprint

Invariance
Interpolation
Adaptive boosting
Bins
Learning algorithms
Feature extraction
Aspect ratio
Classifiers
Pixels
Detectors

All Science Journal Classification (ASJC) codes

  • Software
  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Cite this

@article{e883b49149144788a3e0838bff676133,
title = "Pedestrian detection system using cascaded boosting with invariance of oriented gradients",
abstract = "This study presents a novel learning-based pedestrian detection system capable of automatically detecting individuals of different sizes and orientations against a wide variety of backgrounds, including crowds, even when the individual is partially occluded. To render the detection performance robust toward the effects of geometric and rotational variations in the original image, the feature extraction process is performed using both rectangular- and circular-type blocks of various sizes and aspect ratios. The extracted blocks are rotated in accordance with their dominant orientation(s) such that all the blocks extracted from the input images are rotationally invariant. The pixels within the cells in each block are then voted into rectangular- and circular-type 9-bin histograms of oriented gradients (HOGs) in accordance with their gradient magnitudes and corresponding multivariate Gaussian-weighted windows. Finally, four cell-based histograms are concatenated using a tri-linear interpolation technique to form one 36-dimensional normalized HOG feature vector for each block. The experimental results show that the use of the Gaussian-weighted window approach and tri-linear interpolation technique in constructing the HOG feature vectors improves the detection performance from 91{\%} to 94.5{\%}. In the proposed scheme, the detection process is performed using a cascaded detector structure in which the weak classifiers and corresponding weights of each stage are established using the AdaBoost self-learning algorithm. The experimental results reveal that the cascaded structure not only provides a better detection performance than many of the schemes presented in the literature, but also achieves a significant reduction in the computational time required to classify each input image.",
author = "Wang, {Chi Chen Raxle} and Wu, {Jin Yi} and Lien, {James Jenn-Jier}",
year = "2009",
month = "6",
day = "1",
doi = "10.1142/S0218001409007363",
language = "English",
volume = "23",
pages = "801--823",
journal = "International Journal of Pattern Recognition and Artificial Intelligence",
issn = "0218-0014",
publisher = "World Scientific Publishing Co. Pte Ltd",
number = "4",

}

Pedestrian detection system using cascaded boosting with invariance of oriented gradients. / Wang, Chi Chen Raxle; Wu, Jin Yi; Lien, James Jenn-Jier.

In: International Journal of Pattern Recognition and Artificial Intelligence, Vol. 23, No. 4, 01.06.2009, p. 801-823.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Pedestrian detection system using cascaded boosting with invariance of oriented gradients

AU - Wang, Chi Chen Raxle

AU - Wu, Jin Yi

AU - Lien, James Jenn-Jier

PY - 2009/6/1

Y1 - 2009/6/1

N2 - This study presents a novel learning-based pedestrian detection system capable of automatically detecting individuals of different sizes and orientations against a wide variety of backgrounds, including crowds, even when the individual is partially occluded. To render the detection performance robust toward the effects of geometric and rotational variations in the original image, the feature extraction process is performed using both rectangular- and circular-type blocks of various sizes and aspect ratios. The extracted blocks are rotated in accordance with their dominant orientation(s) such that all the blocks extracted from the input images are rotationally invariant. The pixels within the cells in each block are then voted into rectangular- and circular-type 9-bin histograms of oriented gradients (HOGs) in accordance with their gradient magnitudes and corresponding multivariate Gaussian-weighted windows. Finally, four cell-based histograms are concatenated using a tri-linear interpolation technique to form one 36-dimensional normalized HOG feature vector for each block. The experimental results show that the use of the Gaussian-weighted window approach and tri-linear interpolation technique in constructing the HOG feature vectors improves the detection performance from 91% to 94.5%. In the proposed scheme, the detection process is performed using a cascaded detector structure in which the weak classifiers and corresponding weights of each stage are established using the AdaBoost self-learning algorithm. The experimental results reveal that the cascaded structure not only provides a better detection performance than many of the schemes presented in the literature, but also achieves a significant reduction in the computational time required to classify each input image.

AB - This study presents a novel learning-based pedestrian detection system capable of automatically detecting individuals of different sizes and orientations against a wide variety of backgrounds, including crowds, even when the individual is partially occluded. To render the detection performance robust toward the effects of geometric and rotational variations in the original image, the feature extraction process is performed using both rectangular- and circular-type blocks of various sizes and aspect ratios. The extracted blocks are rotated in accordance with their dominant orientation(s) such that all the blocks extracted from the input images are rotationally invariant. The pixels within the cells in each block are then voted into rectangular- and circular-type 9-bin histograms of oriented gradients (HOGs) in accordance with their gradient magnitudes and corresponding multivariate Gaussian-weighted windows. Finally, four cell-based histograms are concatenated using a tri-linear interpolation technique to form one 36-dimensional normalized HOG feature vector for each block. The experimental results show that the use of the Gaussian-weighted window approach and tri-linear interpolation technique in constructing the HOG feature vectors improves the detection performance from 91% to 94.5%. In the proposed scheme, the detection process is performed using a cascaded detector structure in which the weak classifiers and corresponding weights of each stage are established using the AdaBoost self-learning algorithm. The experimental results reveal that the cascaded structure not only provides a better detection performance than many of the schemes presented in the literature, but also achieves a significant reduction in the computational time required to classify each input image.

UR - http://www.scopus.com/inward/record.url?scp=67650695226&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=67650695226&partnerID=8YFLogxK

U2 - 10.1142/S0218001409007363

DO - 10.1142/S0218001409007363

M3 - Article

VL - 23

SP - 801

EP - 823

JO - International Journal of Pattern Recognition and Artificial Intelligence

JF - International Journal of Pattern Recognition and Artificial Intelligence

SN - 0218-0014

IS - 4

ER -