Bidirectional local template patterns

An effective and discriminative feature for pedestrian detection

Jiu Xu, Ning Jiang, Satoshi Goto

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

In this paper, a novel feature named bidirectional local template patterns (B-LTP) is proposed for use in pedestrian detection in still images. B-LTP is a combination and modification of two features, histogram of templates (HOT) and center-symmetric local binary patterns (CS-LBP). For each pixel, B-LTP defines four templates, each of which contains the pixel itself and two neighboring center-symmetric pixels. For each template, it then calculates information from the relationships among these three pixels and from the two directional transitions across these pixels. Moreover, because the feature length of B-LTP is small, it consumes less memory and computational power. Experimental results on an INRIA dataset show that the speed and detection rate of our proposed B-LTP feature outperform those of other features such as histogram of orientated gradient (HOG), HOT, and covariance matrix (COV).

Original languageEnglish
Pages (from-to)1204-1213
Number of pages10
JournalIEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
VolumeE96-A
Issue number6
DOIs
Publication statusPublished - 2013 Jun

Fingerprint

Pedestrian Detection
Template
Pixels
Pixel
Histogram
Covariance matrix
Data storage equipment
Binary
Gradient
Calculate

Keywords

  • Bidirectional local template patterns
  • Feature extraction
  • Pedestrian detection
  • Support vector machine

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Computer Graphics and Computer-Aided Design
  • Applied Mathematics
  • Signal Processing

Cite this

Bidirectional local template patterns : An effective and discriminative feature for pedestrian detection. / Xu, Jiu; Jiang, Ning; Goto, Satoshi.

In: IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences, Vol. E96-A, No. 6, 06.2013, p. 1204-1213.

Research output: Contribution to journalArticle

@article{d613abb573ce433e9771ea0c9c6d5cc6,
title = "Bidirectional local template patterns: An effective and discriminative feature for pedestrian detection",
abstract = "In this paper, a novel feature named bidirectional local template patterns (B-LTP) is proposed for use in pedestrian detection in still images. B-LTP is a combination and modification of two features, histogram of templates (HOT) and center-symmetric local binary patterns (CS-LBP). For each pixel, B-LTP defines four templates, each of which contains the pixel itself and two neighboring center-symmetric pixels. For each template, it then calculates information from the relationships among these three pixels and from the two directional transitions across these pixels. Moreover, because the feature length of B-LTP is small, it consumes less memory and computational power. Experimental results on an INRIA dataset show that the speed and detection rate of our proposed B-LTP feature outperform those of other features such as histogram of orientated gradient (HOG), HOT, and covariance matrix (COV).",
keywords = "Bidirectional local template patterns, Feature extraction, Pedestrian detection, Support vector machine",
author = "Jiu Xu and Ning Jiang and Satoshi Goto",
year = "2013",
month = "6",
doi = "10.1587/transfun.E96.A.1204",
language = "English",
volume = "E96-A",
pages = "1204--1213",
journal = "IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences",
issn = "0916-8508",
publisher = "Maruzen Co., Ltd/Maruzen Kabushikikaisha",
number = "6",

}

TY - JOUR

T1 - Bidirectional local template patterns

T2 - An effective and discriminative feature for pedestrian detection

AU - Xu, Jiu

AU - Jiang, Ning

AU - Goto, Satoshi

PY - 2013/6

Y1 - 2013/6

N2 - In this paper, a novel feature named bidirectional local template patterns (B-LTP) is proposed for use in pedestrian detection in still images. B-LTP is a combination and modification of two features, histogram of templates (HOT) and center-symmetric local binary patterns (CS-LBP). For each pixel, B-LTP defines four templates, each of which contains the pixel itself and two neighboring center-symmetric pixels. For each template, it then calculates information from the relationships among these three pixels and from the two directional transitions across these pixels. Moreover, because the feature length of B-LTP is small, it consumes less memory and computational power. Experimental results on an INRIA dataset show that the speed and detection rate of our proposed B-LTP feature outperform those of other features such as histogram of orientated gradient (HOG), HOT, and covariance matrix (COV).

AB - In this paper, a novel feature named bidirectional local template patterns (B-LTP) is proposed for use in pedestrian detection in still images. B-LTP is a combination and modification of two features, histogram of templates (HOT) and center-symmetric local binary patterns (CS-LBP). For each pixel, B-LTP defines four templates, each of which contains the pixel itself and two neighboring center-symmetric pixels. For each template, it then calculates information from the relationships among these three pixels and from the two directional transitions across these pixels. Moreover, because the feature length of B-LTP is small, it consumes less memory and computational power. Experimental results on an INRIA dataset show that the speed and detection rate of our proposed B-LTP feature outperform those of other features such as histogram of orientated gradient (HOG), HOT, and covariance matrix (COV).

KW - Bidirectional local template patterns

KW - Feature extraction

KW - Pedestrian detection

KW - Support vector machine

UR - http://www.scopus.com/inward/record.url?scp=84878593078&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84878593078&partnerID=8YFLogxK

U2 - 10.1587/transfun.E96.A.1204

DO - 10.1587/transfun.E96.A.1204

M3 - Article

VL - E96-A

SP - 1204

EP - 1213

JO - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

JF - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

SN - 0916-8508

IS - 6

ER -