ZF/OpenFace
1
0
Fork 0
mirror of https://gitcode.com/gh_mirrors/ope/OpenFace.git synced 2026-05-15 03:37:49 +00:00
Code Issues Packages Projects Releases Wiki Activity

Decoupling C++ code so that AU detection works properly with a new PDM.

Browse Source
This commit is contained in:
Tadas Baltrusaitis
2017-06-20 14:55:59 -04:00
parent 20535fb897
commit 165a708796
45 changed files with 5309 additions and 3039 deletions
Download Patch File Download Diff File Expand all files Collapse all files

104
lib/local/LandmarkDetector/src/LandmarkDetectorFunc.cpp
View File

@@ -71,72 +71,20 @@
using namespace LandmarkDetector;
// Getting a head pose estimate from the currently detected landmarks (rotation with respect to point camera)
// Getting a head pose estimate from the currently detected landmarks, with appropriate correction due to the PDM assuming an orthographic camera
// which is only correct close to the centre of the image
// This method returns a corrected pose estimate with respect to world coordinates with camera at origin (0,0,0)
// The format returned is [Tx, Ty, Tz, Eul_x, Eul_y, Eul_z]
cv::Vec6d LandmarkDetector::GetPoseCamera(const CLNF& clnf_model, double fx, double fy, double cx, double cy)
cv::Vec6d LandmarkDetector::GetPose(const CLNF& clnf_model, double fx, double fy, double cx, double cy)
{
if(!clnf_model.detected_landmarks.empty() && clnf_model.params_global[0] != 0)
{
double Z = fx / clnf_model.params_global[0];
double X = ((clnf_model.params_global[4] - cx) * (1.0/fx)) * Z;
double Y = ((clnf_model.params_global[5] - cy) * (1.0/fy)) * Z;
return cv::Vec6d(X, Y, Z, clnf_model.params_global[1], clnf_model.params_global[2], clnf_model.params_global[3]);
}
else
{
return cv::Vec6d(0,0,0,0,0,0);
}
}
// Getting a head pose estimate from the currently detected landmarks (rotation in world coordinates)
// The format returned is [Tx, Ty, Tz, Eul_x, Eul_y, Eul_z]
cv::Vec6d LandmarkDetector::GetPoseWorld(const CLNF& clnf_model, double fx, double fy, double cx, double cy)
{
if(!clnf_model.detected_landmarks.empty() && clnf_model.params_global[0] != 0)
{
double Z = fx / clnf_model.params_global[0];
double X = ((clnf_model.params_global[4] - cx) * (1.0/fx)) * Z;
double Y = ((clnf_model.params_global[5] - cy) * (1.0/fy)) * Z;
// Here we correct for the camera orientation, for this need to determine the angle the camera makes with the head pose
double z_x = cv::sqrt(X * X + Z * Z);
double eul_x = atan2(Y, z_x);
double z_y = cv::sqrt(Y * Y + Z * Z);
double eul_y = -atan2(X, z_y);
cv::Matx33d camera_rotation = LandmarkDetector::Euler2RotationMatrix(cv::Vec3d(eul_x, eul_y, 0));
cv::Matx33d head_rotation = LandmarkDetector::AxisAngle2RotationMatrix(cv::Vec3d(clnf_model.params_global[1], clnf_model.params_global[2], clnf_model.params_global[3]));
cv::Matx33d corrected_rotation = camera_rotation.t() * head_rotation;
cv::Vec3d euler_corrected = LandmarkDetector::RotationMatrix2Euler(corrected_rotation);
return cv::Vec6d(X, Y, Z, euler_corrected[0], euler_corrected[1], euler_corrected[2]);
}
else
{
return cv::Vec6d(0,0,0,0,0,0);
}
}
// Getting a head pose estimate from the currently detected landmarks, with appropriate correction due to orthographic camera issue
// This is because rotation estimate under orthographic assumption is only correct close to the centre of the image
// This method returns a corrected pose estimate with respect to world coordinates (Experimental)
// The format returned is [Tx, Ty, Tz, Eul_x, Eul_y, Eul_z]
cv::Vec6d LandmarkDetector::GetCorrectedPoseWorld(const CLNF& clnf_model, double fx, double fy, double cx, double cy)
{
if(!clnf_model.detected_landmarks.empty() && clnf_model.params_global[0] != 0)
if (!clnf_model.detected_landmarks.empty() && clnf_model.params_global[0] != 0)
{
// This is used as an initial estimate for the iterative PnP algorithm
double Z = fx / clnf_model.params_global[0];
double X = ((clnf_model.params_global[4] - cx) * (1.0/fx)) * Z;
double Y = ((clnf_model.params_global[5] - cy) * (1.0/fy)) * Z;
double X = ((clnf_model.params_global[4] - cx) * (1.0 / fx)) * Z;
double Y = ((clnf_model.params_global[5] - cy) * (1.0 / fy)) * Z;
// Correction for orientation
// 2D points
@@ -154,35 +102,35 @@ cv::Vec6d LandmarkDetector::GetCorrectedPoseWorld(const CLNF& clnf_model, double
// The camera matrix
cv::Matx33d camera_matrix(fx, 0, cx, 0, fy, cy, 0, 0, 1);
cv::Vec3d vec_trans(X, Y, Z);
cv::Vec3d vec_rot(clnf_model.params_global[1], clnf_model.params_global[2], clnf_model.params_global[3]);
cv::solvePnP(landmarks_3D, landmarks_2D, camera_matrix, cv::Mat(), vec_rot, vec_trans, true);
cv::Vec3d euler = LandmarkDetector::AxisAngle2Euler(vec_rot);
return cv::Vec6d(vec_trans[0], vec_trans[1], vec_trans[2], vec_rot[0], vec_rot[1], vec_rot[2]);
return cv::Vec6d(vec_trans[0], vec_trans[1], vec_trans[2], euler[0], euler[1], euler[2]);
}
else
{
return cv::Vec6d(0,0,0,0,0,0);
return cv::Vec6d(0, 0, 0, 0, 0, 0);
}
}
// Getting a head pose estimate from the currently detected landmarks, with appropriate correction due to perspective projection
// This method returns a corrected pose estimate with respect to a point camera (NOTE not the world coordinates) (Experimental)
// This method returns a corrected pose estimate with respect to a point camera (NOTE not the world coordinates), which is useful to find out if the person is looking at a camera
// The format returned is [Tx, Ty, Tz, Eul_x, Eul_y, Eul_z]
cv::Vec6d LandmarkDetector::GetCorrectedPoseCamera(const CLNF& clnf_model, double fx, double fy, double cx, double cy)
cv::Vec6d LandmarkDetector::GetPoseWRTCamera(const CLNF& clnf_model, double fx, double fy, double cx, double cy)
{
if(!clnf_model.detected_landmarks.empty() && clnf_model.params_global[0] != 0)
if (!clnf_model.detected_landmarks.empty() && clnf_model.params_global[0] != 0)
{
double Z = fx / clnf_model.params_global[0];
double X = ((clnf_model.params_global[4] - cx) * (1.0/fx)) * Z;
double Y = ((clnf_model.params_global[5] - cy) * (1.0/fy)) * Z;
double X = ((clnf_model.params_global[4] - cx) * (1.0 / fx)) * Z;
double Y = ((clnf_model.params_global[5] - cy) * (1.0 / fy)) * Z;
// Correction for orientation
// 3D points
@@ -193,17 +141,17 @@ cv::Vec6d LandmarkDetector::GetCorrectedPoseCamera(const CLNF& clnf_model, doubl
// 2D points
cv::Mat_<double> landmarks_2D = clnf_model.detected_landmarks;
landmarks_2D = landmarks_2D.reshape(1, 2).t();
// Solving the PNP model
// The camera matrix
cv::Matx33d camera_matrix(fx, 0, cx, 0, fy, cy, 0, 0, 1);
cv::Vec3d vec_trans(X, Y, Z);
cv::Vec3d vec_rot(clnf_model.params_global[1], clnf_model.params_global[2], clnf_model.params_global[3]);
cv::solvePnP(landmarks_3D, landmarks_2D, camera_matrix, cv::Mat(), vec_rot, vec_trans, true);
// Here we correct for the camera orientation, for this need to determine the angle the camera makes with the head pose
@@ -219,12 +167,12 @@ cv::Vec6d LandmarkDetector::GetCorrectedPoseCamera(const CLNF& clnf_model, doubl
cv::Matx33d corrected_rotation = camera_rotation * head_rotation;
cv::Vec3d euler_corrected = LandmarkDetector::RotationMatrix2Euler(corrected_rotation);
return cv::Vec6d(vec_trans[0], vec_trans[1], vec_trans[2], euler_corrected[0], euler_corrected[1], euler_corrected[2]);
}
else
{
return cv::Vec6d(0,0,0,0,0,0);
return cv::Vec6d(0, 0, 0, 0, 0, 0);
}
}