The first real-time face detector was implemented using Haar-like feature descriptors by Viola-Jones. The following example is taken from the scikit-image's example that illustrates the extraction, selection, and classification of Haar-like features to detect faces versus non-faces. Clearly, this is a binary classification problem.
The scikit-learn library for feature selection and classification. Although the original implementation for face detection used the AdaBoost ensemble classifier, in this example, a different ensemble classifier, random forest, will be used, primarily to find important Haar-like features useful for the classification.
In order to extract the Haar-like ...