Face feature point detection apparatus and method of the same

A face image feature point detecting method comprising;

• entering an image including a face;
• setting a search area for detecting a target feature point from the image;
• detecting image feature points from the search area of the image;
• storing in advance a first dictionary including information relating peripheral patterns of learning image feature points obtained from a plurality of learning face images and coordinates of relative positions between the respective learning image feature points and learning target feature points associated with respect to each other;
• selecting the peripheral pattern of the learning image feature point which is similar to a peripheral pattern of the image feature point from the peripheral patterns of a plurality of the learning image feature points in the first dictionary;
• extracting a coordinate of the relative position of the learning target feature point corresponding to the selected learning image feature point;
• cropping a peripheral pattern of the target feature point from the face image on the basis of the extracted coordinate of the relative position;
• storing in advance a second dictionary including information relating a peripheral pattern of the learning target feature point obtained from the plurality of learning face images;
• selecting the peripheral pattern of the learning target feature point similar to the peripheral pattern of the target feature point from the peripheral patterns of the plurality of learning target feature points in the second dictionary;
• and
• obtaining the position of the target feature point from the similarity between the peripheral pattern of the image feature point and the peripheral pattern of the selected learning image feature point and the similarity between the peripheral pattern of the target feature point and the peripheral pattern of the selected learning target feature point.

A program stored in a computer readable medium, the program comprising functions of;

• entering an image including a face;
• setting a search area for detecting a target feature point from the image;
• detecting image feature points from the search area of the image;
• storing in advance a first dictionary including information relating peripheral patterns of learning image feature points obtained from a plurality of learning face images and coordinates of relative positions between the respective learning image feature points and learning target feature points associated with respect to each other;
• selecting the peripheral pattern of the learning image feature point which is similar to a peripheral pattern of the image feature point from the peripheral patterns of a plurality of the learning image feature points in the first dictionary;
• extracting a coordinate of the relative position of the learning target feature point corresponding to the selected learning image feature point;
• cropping a peripheral pattern of the target feature point from the face image on the basis of the extracted coordinate of the relative position;
• storing in advance a second dictionary including information relating a peripheral pattern of the learning target feature point obtained from the plurality of learning face images;
• selecting the peripheral pattern of the learning target feature point similar to the peripheral pattern of the target feature point from the peripheral patterns of the plurality of learning target feature points in the second dictionary;
• and
• obtaining the position of the target feature point from the similarity between the peripheral pattern of the image feature point and the peripheral pattern of the selected learning image feature point and the similarity between the peripheral pattern of the target feature point and the peripheral pattern of the selected learning target feature point.