English

GAdaBoost: Accelerating Adaboost Feature Selection with Genetic Algorithms

Computer Vision and Pattern Recognition 2016-09-21 v1

Abstract

Boosted cascade of simple features, by Viola and Jones, is one of the most famous object detection frameworks. However, it suffers from a lengthy training process. This is due to the vast features space and the exhaustive search nature of Adaboost. In this paper we propose GAdaboost: a Genetic Algorithm to accelerate the training procedure through natural feature selection. Specifically, we propose to limit Adaboost search within a subset of the huge feature space, while evolving this subset following a Genetic Algorithm. Experiments demonstrate that our proposed GAdaboost is up to 3.7 times faster than Adaboost. We also demonstrate that the price of this speedup is a mere decrease (3%, 4%) in detection accuracy when tested on FDDB benchmark face detection set, and Caltech Web Faces respectively.

Keywords

Cite

@article{arxiv.1609.06260,
  title  = {GAdaBoost: Accelerating Adaboost Feature Selection with Genetic Algorithms},
  author = {Mai Tolba and Mohamed Moustafa},
  journal= {arXiv preprint arXiv:1609.06260},
  year   = {2016}
}

Comments

8th International Conference on Evolutionary Computation Theory and Applications (ECTA 2016). Final paper will appear at the SCITEPRESS Digital Library

R2 v1 2026-06-22T15:55:43.888Z