Abstract: A feature value extraction apparatus is provided that includes an image input unit that receives an input of an image, and a converted image generation unit that converts the input image into at least one converted image for extracting a feature value of the input image. The apparatus also includes a feature value generation pattern holding unit that stores a feature value generation pattern that includes a plurality of comparison pairs of unit positions for comparing unit values in a unit array in which units, each of which includes at least one pixel, are arranged. The apparatus further includes a feature value extraction unit that compares the unit values corresponding to the comparison pairs through referring to the feature value generation pattern with respect to the converted image, and extracts the feature value of the converted image.

Abstract: A plurality of block pairs are set by setting two of the rectangular blocks contained in one template as one block pair. A score showing a degree to which a part of an object is included in the template is held based on a relation of a relative level between block luminance values in each of the block pairs set in the template. These processes are carried out for all templates. Based on a sum of the scores of all the templates, and whether the object is included in the image is determined.

Abstract: A feature value extraction apparatus includes: an image input unit that receives an input of an image; a converted image generation unit that converts the input image into at least one converted image for extracting a feature value of the input image; a feature value generation pattern holding unit that stores a feature value generation pattern that includes a plurality of comparison pairs of unit positions for comparing unit values in a unit array in which units, each of which includes at least one pixel, are arranged; and a feature value extraction unit that compares the unit values corresponding to the comparison pairs through referring to the feature value generation pattern with respect to the converted image, and extracts the feature value of the converted image.

Abstract: A plurality of block pairs are set by setting two of the rectangular blocks contained in one template as one block pair. A score showing a degree to which a part of an object is included in the template is held based on a relation of a relative level between block luminance values in each of the block pairs set in the template. These processes are carried out for all templates. Based on a sum of the scores of all the templates, and whether the object is included in the image is determined.