A Binary Decision Tree Implementation of a Boosted Strong Classifier

Shaohua Kevin Zhou

Abstract

Viola and Jones [1] proposed the influential rapid object detection algorithm. They used AdaBoost to select from a large pool a set of simple features and constructed a strong classifier of the form {∑ j α j h j (x) ≥ θ} where each h j (x) is a binary weak classifier based on a simple feature. In this paper, we construct, using statistical detection theory, a binary decision tree from the strong classifier of the above form. Each node of the decision tree is just a weak classifier and the knowledge of the coefficients α j is no longer needed. Also, the binary tree has a lot of early exits. As a result, we achieve an automatic speedup that always makes the rapid Viola and Jones algorithm rapider.
BibTeX
@inproceedings{zhou2005binary,
  author = {Shaohua Kevin Zhou},
  title = {{A Binary Decision Tree Implementation of a Boosted Strong Classifier}},
  booktitle = {AMFG},
  year = {2005},
  doi = {10.1007/11564386\_16},
}

Loading related content...

Figure