An image input unit configured to enter a face image, a search area and scale setting unit configured to set a search area and a scale, an image feature point detection unit configured to detect image feature points selected from local image information of respective points, a first dictionary storing coordinates of the relative position between the image feature points and a target feature point in association with peripheral patterns of the image feature points, a first pattern matching unit configured to match the first dictionary and the peripheral patterns of the image feature points, a target feature point candidate position estimating unit configured to estimate candidates of the position of the target feature point, a second dictionary, a second pattern matching unit configured to match the second dictionary and a peripheral pattern of the target feature point and a determination unit configured to obtain the position of the target feature point.