Cascading classifiers

Cascading is a particular case of ensemble learning based on the concatenation of several classifiers, using all information collected from the output from a given classifier as additional information for the next classifier in the cascade. Unlike voting or stacking ensembles, which are multiexpert systems, cascading is a multistage one. Cascading classifiers are trained with several hundred "positive" sample views of a particular object and arbitrary "negative" images of the same size. After the classifier is trained it can be applied to a region of an image and detect the object in question. To search for the object in the entire frame, the search window can be moved across the image and check every location with the classifier. This process is most commonly used in for object detection and tracking, primarily facial detection and recognition. The first cascading classifier was the face detector of Viola and Jones (2001). The requirement for this classifier was to be fast in order to be implemented on low-power CPUs, such as cameras and phones. It can be seen from this description that the classifier will not accept faces that are upside down (the eyebrows are not in a correct position) or the side of the face (the nose is no longer in the center, and shadows on the side of the nose might be missing). Separate cascade classifiers have to be trained for every rotation that is not in the image plane (side of face) and will have to be retrained or run on rotated features for every rotation that is in the image plane (face upside down or tilted to the side). Scaling is not a problem, since the features can be scaled (centerpixel, leftpixels and rightpixels have a dimension only relative to the rectangle examined). In recent cascades, pixel value from some part of a rectangle compared to another have been replaced with Haar wavelets. To have good overall performance, the following criteria must be met: Each stage must validate all faces, and can produce many false positives.