The Cascade of Classifiers

The performance of one classifier is not suitable for object classification, since it produces a high hit rate, e.g., 0.999, and error rate, e.g., 0.5. Nevertheless the hit rate is much higher than the error rate. To construct an overall good classifier, several classifiers are arranged in a cascade, i.e., a degenerated decision tree. In every stage of the cascade a decision is made whether the image contains the object or not. This computation reduces both rates. Since the hit rate is close to one, their multiplication results also in a value close to one, while the multiplication of the smaller error rates approaches zero. Furthermore the whole classification process speeds up. Figure 3 shows an example cascade of classifiers for detecting an ``office chair'' in depth images.

Figure: The first three stages of a cascade of classifiers to detect an office chair in depth data. Every stage contains several simple classifiers that use Haar-like features.

An overall effective cascade is learned by a simple iterative method. For every stage the classification function $h(x)$ is learned, until the required hit rate is reached. The process continues with the next stage using only the currently misclassified negative examples. The number of features used in each classifiers increases with additional stages (Figure 3).