How to get Euler angles?

[chen1169968275]

No description provided.

[chen1169968275]

As your code in HeadPoseEstimation::pose(), after you get the rvec or pose matrix , how to get the face real Euler angles?(the value of yaw, pitch and roll)

[alexis-jacq]

Thankyou for your interest !

We use ROS TF library to do it. You can see in src/ros_head_pose_estimator.cpp the following lines :

        auto poses = estimator.poses();
[...]
        tf::Quaternion qrot;
        tf::Matrix3x3 mrot(
                trans(0,0), trans(0,1), trans(0,2),
                trans(1,0), trans(1,1), trans(1,2),
                trans(2,0), trans(2,1), trans(2,2));
        mrot.getRotation(qrot);
        face_pose.setRotation(qrot);

qrot is a quaternion, if you want euler angles you can use the matrix mrot as follow:

double roll, pitch, yaw;
mrot.getRPY(roll, pitch, yaw);

Please also have a look to our new package OpenFace Tracker, the head pose estimation is much more robust !

[chen1169968275]

Thanks for your reply.
With ROS TF Library, as your codes, I can get the Euler angles.
But there is another problem. If I cannot use ROS TF library, we get the rvec matrix after solvePnP() function, I don't know weather the rvec matrix is the 3D rotation matrix or not. If so, can I get the Euler angles with OPENCV's api decomposeProjectionMatrix to get the Euler angles?

[alexis-jacq]

rvec is a rotation vector (output of solvePNP), we use an OPENCV function "Rodigues" to get the 3x3 rotation matrix given this vector.
Then, you can use OPENCV's decomposeProjectionMatrix to get Euler angles.

[chen1169968275]

Thanks.
So I do as follows:

cv::Mat rvec, tvec;
solvePnP(head_points, detected_points, projection, cv:;noArray(), rvec, tvec, false, cv::ITERATIVE);
cv::Mat rodrigues_rvec;
cv::Rodrigues(rvec, rodrigues_rvec);
cv::Mat cameraMatrix, rotMatrix, transVect, rotMatrixX, rotMatrixY, rotMatrixZ;
cv::Vec3d eulerAngles;
double* _r = rodrigues_rvec.ptr();
double projMatrix[12] = {_r[0], _r[1], _r[2], 0, _r[3], _r[4], _r[5], 0, _r[6], _r[7], _r[8], _r[9], 0};
cv::decomposeProjectionMatrix(cv::Mat(3,4,CV_64FC1,projMatrix, cameraMatrix, rotMatrix, transVect, rotMatrixX, rotMatrixY, rotMatrixZ, eulerAngles);

After that , the 3 value of eulerAngles can reach 100+ or smaller than -100, I cannot understand what does the values mean.
If possible, can you update your samples/head_pose_estimation_test.cpp with show the face's Euler angles.

[KeyKy]

@chen1169968275 do you solved this?