A modular approach to training cascades of boosted ensembles

Teo Susnjak*, Andre L. Barczak, Ken A. Hawick

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

2 Citations (Scopus)

Abstract

Building on the ideas of Viola-Jones [1] we present a framework for training cascades of boosted ensembles (CoBE) which introduces further modularity and tractability to the training process. It addresses the challenges faced by CoBE frameworks such as protracted runtimes, slow layer convergences and classifier optimization. The framework possesses the ability to bootstrap positive samples and may in turn be extended into the domain of incremental learning. This paper aims to address our framework's susceptibility to overfitting with possible solutions. Experiments are conducted on face detectors using the bootstrapping of large positive datasets and their accuracy, with respect to overfitting, is examined.

Original languageEnglish
Title of host publicationStructural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, SSPR and SPR 2010, Proceedings
EditorsEdwin R. Hancock, Richard C. Wilson, Terry Windeatt, Ikay Ulusoy, Francisco Escolano
PublisherSpringer
Pages640-649
Number of pages10
ISBN (Print)3642149790, 9783642149795
DOIs
Publication statusPublished - 2010
Externally publishedYes
Event7th Joint IAPR International Workshop on Structural, Syntactic, and Statistical Pattern Recognition, SSPR and SPR 2010 - Cesme, Izmir, Turkey
Duration: 18 Aug 201020 Aug 2010

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume6218 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Workshop

Workshop7th Joint IAPR International Workshop on Structural, Syntactic, and Statistical Pattern Recognition, SSPR and SPR 2010
Country/TerritoryTurkey
CityCesme, Izmir
Period18/08/1020/08/10

Fingerprint

Dive into the research topics of 'A modular approach to training cascades of boosted ensembles'. Together they form a unique fingerprint.

Cite this