As seen in the previous section 4.4.2, OpenCV’s face detector uses the Viola- Jones method for detecting an object which consists in Haar-like features that are more exactly simple rectangular features, an Integral Image for rapid feature detection, a machine-learning method, the AdaBoost, and a cascade classifier to efficiently combine multiple features. Haar features are Haar wavelets that are single wavelength square waves with one high interval and one low interval. In two dimensions, a square wave is a pair of adjacent rectangles, one light and one dark. The actual rectangle combinations
used for the visual detection of objects in OpenCV are not true Haar wavelets, but instead contain rectangle combinations better suited to visual recognition tasks. Due to this difference, the features are called Haar-like features and not Haar wavelets, and can be seen in Figure 4.12.