@inproceedings{89de61c2872a4d0aa0870299c1e3d936,
title = "A modular approach to training cascades of boosted ensembles",
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.",
author = "Teo Susnjak and Barczak, {Andre L.} and Hawick, {Ken A.}",
year = "2010",
doi = "10.1007/978-3-642-14980-1_63",
language = "English",
isbn = "3642149790",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "640--649",
editor = "Hancock, {Edwin R.} and Wilson, {Richard C.} and Terry Windeatt and Ikay Ulusoy and Francisco Escolano",
booktitle = "Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, SSPR and SPR 2010, Proceedings",
address = "Germany",
note = "7th Joint IAPR International Workshop on Structural, Syntactic, and Statistical Pattern Recognition, SSPR and SPR 2010 ; Conference date: 18-08-2010 Through 20-08-2010",
}