Reinforced AdaBoost learning for object detection with local pattern representations

Younghyun Lee, David K. Han, Hanseok Ko

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

A reinforced AdaBoost learning algorithm is proposed for object detection with local pattern representations. In implementing Adaboost learning, the proposed algorithm employs an exponential criterion as a cost function and Newton's method for its optimization. In particular, we introduce an optimal selection of weak classifiers minimizing the cost function and derive the reinforced predictions based on a judicial confidence estimate to determine the classification results. The weak classifier of the proposed method produces real-valued predictions while that of the conventional Adaboost method produces integer valued predictions of +1 or -1. Hence, in the conventional learning algorithms, the entire sample weights are updated by the same rate. On the contrary, the proposed learning algorithm allows the sample weights to be updated individually depending on the confidence level of each weak classifier prediction, thereby reducing the number of weak classifier iterations for convergence. Experimental classification performance on human face and license plate images confirm that the proposed method requires smaller number of weak classifiers than the conventional learning algorithm, resulting in higher learning and faster classification rates. An object detector implemented based on the proposed learning algorithm yields better performance in field tests in terms of higher detection rate with lower false positives than that of the conventional learning algorithm.

Original languageEnglish
Article number153465
JournalThe Scientific World Journal
Volume2013
DOIs
Publication statusPublished - 2013 Dec 1

Fingerprint

Adaptive boosting
Learning algorithms
learning
Learning
Classifiers
Cost functions
prediction
Newton-Raphson method
Weights and Measures
Costs and Cost Analysis
Object detection
detection
Licensure
cost
Detectors
method

ASJC Scopus subject areas

  • Biochemistry, Genetics and Molecular Biology(all)
  • Environmental Science(all)
  • Medicine(all)

Cite this

Reinforced AdaBoost learning for object detection with local pattern representations. / Lee, Younghyun; Han, David K.; Ko, Hanseok.

In: The Scientific World Journal, Vol. 2013, 153465, 01.12.2013.

Research output: Contribution to journalArticle

@article{11279ca4794f42788846dea042829007,
title = "Reinforced AdaBoost learning for object detection with local pattern representations",
abstract = "A reinforced AdaBoost learning algorithm is proposed for object detection with local pattern representations. In implementing Adaboost learning, the proposed algorithm employs an exponential criterion as a cost function and Newton's method for its optimization. In particular, we introduce an optimal selection of weak classifiers minimizing the cost function and derive the reinforced predictions based on a judicial confidence estimate to determine the classification results. The weak classifier of the proposed method produces real-valued predictions while that of the conventional Adaboost method produces integer valued predictions of +1 or -1. Hence, in the conventional learning algorithms, the entire sample weights are updated by the same rate. On the contrary, the proposed learning algorithm allows the sample weights to be updated individually depending on the confidence level of each weak classifier prediction, thereby reducing the number of weak classifier iterations for convergence. Experimental classification performance on human face and license plate images confirm that the proposed method requires smaller number of weak classifiers than the conventional learning algorithm, resulting in higher learning and faster classification rates. An object detector implemented based on the proposed learning algorithm yields better performance in field tests in terms of higher detection rate with lower false positives than that of the conventional learning algorithm.",
author = "Younghyun Lee and Han, {David K.} and Hanseok Ko",
year = "2013",
month = "12",
day = "1",
doi = "10.1155/2013/153465",
language = "English",
volume = "2013",
journal = "Scientific World Journal",
issn = "2356-6140",
publisher = "Hindawi Publishing Corporation",

}

TY - JOUR

T1 - Reinforced AdaBoost learning for object detection with local pattern representations

AU - Lee, Younghyun

AU - Han, David K.

AU - Ko, Hanseok

PY - 2013/12/1

Y1 - 2013/12/1

N2 - A reinforced AdaBoost learning algorithm is proposed for object detection with local pattern representations. In implementing Adaboost learning, the proposed algorithm employs an exponential criterion as a cost function and Newton's method for its optimization. In particular, we introduce an optimal selection of weak classifiers minimizing the cost function and derive the reinforced predictions based on a judicial confidence estimate to determine the classification results. The weak classifier of the proposed method produces real-valued predictions while that of the conventional Adaboost method produces integer valued predictions of +1 or -1. Hence, in the conventional learning algorithms, the entire sample weights are updated by the same rate. On the contrary, the proposed learning algorithm allows the sample weights to be updated individually depending on the confidence level of each weak classifier prediction, thereby reducing the number of weak classifier iterations for convergence. Experimental classification performance on human face and license plate images confirm that the proposed method requires smaller number of weak classifiers than the conventional learning algorithm, resulting in higher learning and faster classification rates. An object detector implemented based on the proposed learning algorithm yields better performance in field tests in terms of higher detection rate with lower false positives than that of the conventional learning algorithm.

AB - A reinforced AdaBoost learning algorithm is proposed for object detection with local pattern representations. In implementing Adaboost learning, the proposed algorithm employs an exponential criterion as a cost function and Newton's method for its optimization. In particular, we introduce an optimal selection of weak classifiers minimizing the cost function and derive the reinforced predictions based on a judicial confidence estimate to determine the classification results. The weak classifier of the proposed method produces real-valued predictions while that of the conventional Adaboost method produces integer valued predictions of +1 or -1. Hence, in the conventional learning algorithms, the entire sample weights are updated by the same rate. On the contrary, the proposed learning algorithm allows the sample weights to be updated individually depending on the confidence level of each weak classifier prediction, thereby reducing the number of weak classifier iterations for convergence. Experimental classification performance on human face and license plate images confirm that the proposed method requires smaller number of weak classifiers than the conventional learning algorithm, resulting in higher learning and faster classification rates. An object detector implemented based on the proposed learning algorithm yields better performance in field tests in terms of higher detection rate with lower false positives than that of the conventional learning algorithm.

UR - http://www.scopus.com/inward/record.url?scp=84893872147&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84893872147&partnerID=8YFLogxK

U2 - 10.1155/2013/153465

DO - 10.1155/2013/153465

M3 - Article

VL - 2013

JO - Scientific World Journal

JF - Scientific World Journal

SN - 2356-6140

M1 - 153465

ER -