Face image generating method for recognizing face

A face image generating method for recognizing face performed by a face image generating device to extract a face image from image frames collected from an image acquiring device and generate a face image for face recognition, the method comprising:

• a first step of converting an RGB-based image frame received from the image acquiring device into a YCbCr image;
• a second step of setting a region satisfying a skin color condition in the YCbCr image as a face candidate region;
• a third step of determining a face region based on whether at least two feature points corresponding to an eye, a nose, or a mouth is included in the face candidate region;
• a fourth step of classifying a front face or a side face based on an eye feature existing in the face region and an aspect ratio of the face region and extracting each face region, thereby generating a front face image and a side face image;
• a fifth step of calculating a degree of inclination of a side face in the side face image based on location information of a nose and a chin and generating a true side face image using the calculated degree of inclination;
• and
• a sixth step of generating a face image for recognition by converting the front face image and the side face image into different sizes according to a distance to the image acquiring device.