Method of controlling digital image processing apparatus for face detection, and digital image processing apparatus employing the method
View Patent ↗Provided is a method of controlling a digital image processing apparatus for detecting a face from continuously input images, the method comprising operations (a) to (c). In (a), if a face is detected, image information of a body area is stored. In (b), if the face is not detected, a body having the image information stored in (a) is detected. In (c), if a current body is detected after a previous body was detected in (b), an image characteristic of the previously detected body is compared to an image characteristic of the currently detected body, and a movement state of the face is determined according to the comparison result.
1. A method of controlling a digital image processing apparatus for detecting a face from continuously input images, the method comprising:
(a) if a face is detected, storing image information of a body area;
(b) if a face is not detected, detecting a body having image information stored in (a); and
(c) if a current body is detected after a previous body was detected in (b), determining a movement state of the face by comparing a luminance characteristic of contours of the previously detected body and a luminance characteristic of contours of the currently detected body by: (c11) obtaining a contour feature value by adding luminance difference values between adjacent pixels with respect to all contours of the previously detected body; (c12) obtaining a contour feature value by adding luminance difference values between adjacent pixels with respect to all contours of the currently detected body; (c13) calculating a rate of contour feature change, which is a rate of change of the contour feature value of the currently detected body with respect to the contour feature value of the previously detected body; and (c14) determining the movement state of the face according to the rate of contour feature change.
2. The method of claim 1 , wherein (a) comprises:
(a1) setting a size and reference coordinates of a body according to a size and reference coordinates of the detected face; and
(a2) storing image information of the body area set in (a1).
3. The method of claim 1 , wherein, in (a) and (b), the image information of the body area comprises a mean gradation of the body area.
4. The method of claim 2 , wherein the mean gradation of the body area comprises a mean red gradation, a mean green gradation, and a mean blue gradation of the body area.
5. The method of claim 1 , wherein, in (c11) and (c12), when pixels of a contour are a second pixel and a third pixel, a luminance difference value between a first pixel and the second pixel, a luminance difference value between the second pixel and the third pixel, and a luminance difference value between the third pixel and a fourth pixel are added with respect to the contour.
6. The method of claim 1 , wherein in (c), the movement state of the face is determined by comparing a luminance histogram of the previously detected body image and a luminance histogram of the currently detected body image.
7. The method of claim 6 , further comprising:
(c21) obtaining a histogram corresponding to the number of pixels according to corresponding luminance values of the previously detected body image;
(c22) obtaining a histogram corresponding to the number of pixels according to corresponding luminance values of the currently detected body image;
(c23) obtaining a transition of the histogram of the currently detected body image with respect to the histogram of the previously detected body image; and
(c24) determining the movement state of the face according to the histogram transition.
8. A digital image processing apparatus, which comprises a main controller and detects a face from continuously input images, wherein the main controller employs the method of claim 1 .