Multi-objective neural network optimization for visual object detection

Publikation: Bidrag til bog/antologi/rapport › Bidrag til bog/antologi › Forskning › fagfællebedømt

Standard

Multi-objective neural network optimization for visual object detection. / Roth, Stefan; Gepperth, Alexander; Igel, Christian.

Multi-objective machine learning. red. / Yaochu Jin. Bind V 2006. s. 629-655 (Studies in Computational Intelligence, Bind 16).

Publikation: Bidrag til bog/antologi/rapport › Bidrag til bog/antologi › Forskning › fagfællebedømt

Harvard

Roth, S, Gepperth, A & Igel, C 2006, Multi-objective neural network optimization for visual object detection. i Y Jin (red.), Multi-objective machine learning. bind V, Studies in Computational Intelligence, bind 16, s. 629-655. https://doi.org/10.1007/11399346_27

APA

Roth, S., Gepperth, A., & Igel, C. (2006). Multi-objective neural network optimization for visual object detection. I Y. Jin (red.), Multi-objective machine learning (Bind V, s. 629-655). Studies in Computational Intelligence Bind 16 https://doi.org/10.1007/11399346_27

Vancouver

Roth S, Gepperth A, Igel C. Multi-objective neural network optimization for visual object detection. I Jin Y, red., Multi-objective machine learning. Bind V. 2006. s. 629-655. (Studies in Computational Intelligence, Bind 16). https://doi.org/10.1007/11399346_27

Author

Roth, Stefan ; Gepperth, Alexander ; Igel, Christian. / Multi-objective neural network optimization for visual object detection. Multi-objective machine learning. red. / Yaochu Jin. Bind V 2006. s. 629-655 (Studies in Computational Intelligence, Bind 16).

Bibtex

@inbook{8c33f8ffa902441fae696a730ca9d75a,
title = "Multi-objective neural network optimization for visual object detection",
abstract = "In real-time computer vision, there is a need for classifiers that detect patterns fast and reliably. We apply multi-objective optimization (MOO) to the design of feed-forward neural networks for real-world object recognition tasks, where computational complexity and accuracy define partially conflicting objectives. Evolutionary structure optimization and pruning are compared for the adaptation of the network topology. In addition, the results of MOO are contrasted to those of a single-objective evolutionary algorithm. As a part of the evolutionary algorithm, the automatic adaptation of operator probabilities in MOO is described.",
author = "Stefan Roth and Alexander Gepperth and Christian Igel",
year = "2006",
doi = "10.1007/11399346_27",
language = "English",
isbn = "978-3-540-30676-4",
volume = "V",
series = "Studies in Computational Intelligence",
publisher = "Springer",
pages = "629--655",
editor = "Yaochu Jin",
booktitle = "Multi-objective machine learning",

}

RIS

TY - CHAP

T1 - Multi-objective neural network optimization for visual object detection

AU - Roth, Stefan

AU - Gepperth, Alexander

AU - Igel, Christian

PY - 2006

Y1 - 2006

N2 - In real-time computer vision, there is a need for classifiers that detect patterns fast and reliably. We apply multi-objective optimization (MOO) to the design of feed-forward neural networks for real-world object recognition tasks, where computational complexity and accuracy define partially conflicting objectives. Evolutionary structure optimization and pruning are compared for the adaptation of the network topology. In addition, the results of MOO are contrasted to those of a single-objective evolutionary algorithm. As a part of the evolutionary algorithm, the automatic adaptation of operator probabilities in MOO is described.

AB - In real-time computer vision, there is a need for classifiers that detect patterns fast and reliably. We apply multi-objective optimization (MOO) to the design of feed-forward neural networks for real-world object recognition tasks, where computational complexity and accuracy define partially conflicting objectives. Evolutionary structure optimization and pruning are compared for the adaptation of the network topology. In addition, the results of MOO are contrasted to those of a single-objective evolutionary algorithm. As a part of the evolutionary algorithm, the automatic adaptation of operator probabilities in MOO is described.

U2 - 10.1007/11399346_27

DO - 10.1007/11399346_27

M3 - Book chapter

AN - SCOPUS:33845294450

SN - 978-3-540-30676-4

VL - V

T3 - Studies in Computational Intelligence

SP - 629

EP - 655

BT - Multi-objective machine learning

A2 - Jin, Yaochu

ER -

ID: 168323444