Use of simple n-rectangular features in the method of rapid object detection on images

In given article use of simple n-rectangular features for realization of object detection algorithm on images is considered. Among all set of attributes get out a few strongest which are expedient for using at initial stages of search of object for rejection of the big number of background images. Necessity and validity of restriction of the sizes of attributes for decrease influence of noise is considered. An example works of the multistage qualifier on test set is given.