Convolution neural network with selective multi-stage feature fusion

Case study on vehicle rear detection

Won Jae Lee, Dong W. Kim, Tae Koo Kang, Myo Taeg Lim

Research output: Contribution to journalArticle

Abstract

Vision-based vehicle detection is the most basic and important technology in advanced driver assistance systems. In this paper, we propose a vehicle detection framework using selective multi-stage features in convolutional neural networks (CNNs) to improve vehicle detection performance. A 10-layer CNN model was designed and visualization techniques were used to selectively extract features from the activation feature map, called selective multi-stage features. The proposed features contain characteristic vehicle image information and are more robust than traditional features against noise. We trained the AdaBoost algorithmusing these features to implement a vehicle detector. The experimental results verified that the proposed vehicle detection framework exhibited better performance than previous frameworks.

Original languageEnglish
Article number2468
JournalApplied Sciences (Switzerland)
Volume8
Issue number12
DOIs
Publication statusPublished - 2018 Dec 3

Fingerprint

Convolution
convolution integrals
vehicles
Fusion reactions
fusion
Neural networks
Advanced driver assistance systems
Adaptive boosting
Visualization
Chemical activation
activation
Detectors
detectors

Keywords

  • AdaBoost
  • Convolutional neural network
  • Feature extraction
  • Vehicle detection

ASJC Scopus subject areas

  • Materials Science(all)
  • Instrumentation
  • Engineering(all)
  • Process Chemistry and Technology
  • Computer Science Applications
  • Fluid Flow and Transfer Processes

Cite this

Convolution neural network with selective multi-stage feature fusion : Case study on vehicle rear detection. / Lee, Won Jae; Kim, Dong W.; Kang, Tae Koo; Lim, Myo Taeg.

In: Applied Sciences (Switzerland), Vol. 8, No. 12, 2468, 03.12.2018.

Research output: Contribution to journalArticle

@article{65427646b7194e368ad591279e5c5a72,
title = "Convolution neural network with selective multi-stage feature fusion: Case study on vehicle rear detection",
abstract = "Vision-based vehicle detection is the most basic and important technology in advanced driver assistance systems. In this paper, we propose a vehicle detection framework using selective multi-stage features in convolutional neural networks (CNNs) to improve vehicle detection performance. A 10-layer CNN model was designed and visualization techniques were used to selectively extract features from the activation feature map, called selective multi-stage features. The proposed features contain characteristic vehicle image information and are more robust than traditional features against noise. We trained the AdaBoost algorithmusing these features to implement a vehicle detector. The experimental results verified that the proposed vehicle detection framework exhibited better performance than previous frameworks.",
keywords = "AdaBoost, Convolutional neural network, Feature extraction, Vehicle detection",
author = "Lee, {Won Jae} and Kim, {Dong W.} and Kang, {Tae Koo} and Lim, {Myo Taeg}",
year = "2018",
month = "12",
day = "3",
doi = "10.3390/app8122468",
language = "English",
volume = "8",
journal = "Applied Sciences (Switzerland)",
issn = "2076-3417",
publisher = "Multidisciplinary Digital Publishing Institute",
number = "12",

}

TY - JOUR

T1 - Convolution neural network with selective multi-stage feature fusion

T2 - Case study on vehicle rear detection

AU - Lee, Won Jae

AU - Kim, Dong W.

AU - Kang, Tae Koo

AU - Lim, Myo Taeg

PY - 2018/12/3

Y1 - 2018/12/3

N2 - Vision-based vehicle detection is the most basic and important technology in advanced driver assistance systems. In this paper, we propose a vehicle detection framework using selective multi-stage features in convolutional neural networks (CNNs) to improve vehicle detection performance. A 10-layer CNN model was designed and visualization techniques were used to selectively extract features from the activation feature map, called selective multi-stage features. The proposed features contain characteristic vehicle image information and are more robust than traditional features against noise. We trained the AdaBoost algorithmusing these features to implement a vehicle detector. The experimental results verified that the proposed vehicle detection framework exhibited better performance than previous frameworks.

AB - Vision-based vehicle detection is the most basic and important technology in advanced driver assistance systems. In this paper, we propose a vehicle detection framework using selective multi-stage features in convolutional neural networks (CNNs) to improve vehicle detection performance. A 10-layer CNN model was designed and visualization techniques were used to selectively extract features from the activation feature map, called selective multi-stage features. The proposed features contain characteristic vehicle image information and are more robust than traditional features against noise. We trained the AdaBoost algorithmusing these features to implement a vehicle detector. The experimental results verified that the proposed vehicle detection framework exhibited better performance than previous frameworks.

KW - AdaBoost

KW - Convolutional neural network

KW - Feature extraction

KW - Vehicle detection

UR - http://www.scopus.com/inward/record.url?scp=85057842574&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85057842574&partnerID=8YFLogxK

U2 - 10.3390/app8122468

DO - 10.3390/app8122468

M3 - Article

VL - 8

JO - Applied Sciences (Switzerland)

JF - Applied Sciences (Switzerland)

SN - 2076-3417

IS - 12

M1 - 2468

ER -