I am using "facedetect.py" for the detection of face and eyes. Now I want to calculate the angle (angle of elevation) at which face and eyes are detected. Can anyone Please help me in that.
If it is basically a face alignment that you want to achieve, then use this code snippet
double eyeXdis = eye_r.x - eye_l.x;
double eyeYdis = eye_r.y - eye_l.y;
double angle = atan(eyeYdis/eyeXdis);
double degree = angle*180/CV_PI;
double desired_eye_distance = 44.0;
double scale = desired_eye_distance / eyeXdis;
Mat res;
Point2f center(test.cols/2, test.rows/2);
Mat rot = getRotationMatrix2D(center, degree, scale);
warpAffine(test, res, rot, Size(), INTER_CUBIC, BORDER_CONSTANT, Scalar(127));
// probably crop this again from center
Credits to @berak for posting it in another topic this week.
@StevenPuttemans The "test" variable described in your solution is the image matrix, right? It's possible to apply the same solution to a set of points, i.e. If I have an matrix 2xn where each elements in row 1 represent the X coordinate of a vector and each element in row represents a Y coordinate.
Yep that is the image matrix indeed. As far as I know you could indeed get a transformation matrix if you give it 2 sets of points. However this is not done the same way as above. Look at the documentation of warpAffine. It has a sample where you calculate the transformation based on 2 groups of 3 points in space.
Asked: 2015-02-10 12:30:53 -0600
Seen: 2,150 times
Last updated: Feb 11 '15
Can you explain your problem with a sample image?