OBJECT COLLATING METHOD

PURPOSE:To efficiently recognize an object by constituting a retina model based on an image processing function of a man's eye as a reference, and constituting automatically the image feature quantity of the object in the same way as the man's eye. CONSTITUTION:A place in which frequency by which a moving locus of the center of a mask 3 constituted by using plural pieces of blocks consisting of plural picture elements passes through is high and a representative locus pattern are used as the feature quantity, therefore, a feature point of an image is discriminated instantaneously with the same way as a man's retina, and also, by its eyeball moving locus, an object can be recognized. In such a way, by collecting in advance the feature quantity of the image of various objects in a collation dictionary 6, an unknown object in the image can be discriminated efficiently with a high discrimination rate without being influenced by misalignment at the time of photographing and a noise on the image.