Multi-objective neural network optimization for visual object detection

Publikation: Bidrag til bog/antologi/rapportBidrag til bog/antologiForskningfagfællebedømt

Stefan Roth, Alexander Gepperth, Christian Igel

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.

OriginalsprogEngelsk
TitelMulti-objective machine learning
RedaktørerYaochu Jin
Antal sider27
Vol/bindV
Publikationsdato2006
Sider629-655
ISBN (Trykt)978-3-540-30676-4
ISBN (Elektronisk)978-3-540-33019-6
DOI
StatusUdgivet - 2006
Eksternt udgivetJa
NavnStudies in Computational Intelligence
Vol/bind16
ISSN1860-949X

ID: 168323444