Disclosed is a method, apparatus, and system for detecting a human face in an image. The method includes, for a subset of pixels in the image, deriving a first variable from the gray-level distribution of the image and deriving a second variable from a preset reference distribution that is characteristic of the object. The method further includes evaluating the correspondence between the first variable and the second variable over the subset of pixels and determining if the image contains the object based on the result of this evaluation.