Automatic cascade training with perturbation bias

Jie Sun, James M. Rehg, Aaron Bobick

Research output: Contribution to journalConference articlepeer-review


Face detection methods based on a cascade architecture have demonstrated fast and robust performance. Cascade learning is aided by the modularity of the architecture in which nodes are chained together to form a cascade. In this paper we present two new cascade learning results which address the decoupled nature of the cascade learning task. First, we introduce a cascade indifference curve framework which connects the learning objectives for a node to the overall cascade performance. We derive a new cost function for node learning which yields fully-automatic stopping conditions and improved detection performance. Second, we introduce the concept of perturbation bias which leverages the statistical differences between target and non-target classes in a detection problem to obtain improved performance and robustness. We derive necessary and sufficient conditions for the success of the method and present experimental results.

Original languageEnglish (US)
Pages (from-to)II276-II283
JournalProceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
StatePublished - 2004
Externally publishedYes
EventProceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2004 - Washington, DC, United States
Duration: Jun 27 2004Jul 2 2004

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition

Cite this