Where can I find the camera matrix params derived from the model for a given frame?

[Tetsujinfr]

thanks

[Tetsujinfr]

Ok so I guess this is
self._get_camera_matrix = partial(self._interpreter.get_tensor,
output_details[0]["index"])

in BaseTFLiteFaceAlignment ?

[1996scarlet]

Yes, we can get the camera_matrix R directly from the model output.

Dense-Head-Pose-Estimation/service/TFLiteFaceAlignment.py

Lines 29 to 30 in 60a5bdf

	self._get_camera_matrix = partial(self._interpreter.get_tensor,
	output_details[0]["index"])

Dense-Head-Pose-Estimation/service/TFLiteFaceAlignment.py

Lines 75 to 82 in 60a5bdf

	def _postprocessing(self, M):
	iM = cv2.invertAffineTransform(M)
	R = self._get_camera_matrix()[0]
	landmarks = self._decode_landmarks(iM)
	return landmarks, R

Then, we can decompose matrix to ruler angle

Dense-Head-Pose-Estimation/service/__init__.py

Lines 101 to 111 in 60a5bdf

	def pose(frame, results, color):
	landmarks, params = results
	# rotate matrix
	R = params[:3, :3].copy()
	# decompose matrix to ruler angle
	euler = rotationMatrixToEulerAngles(R)
	print(f"Pitch: {euler[0]}; Yaw: {euler[1]}; Roll: {euler[2]};")
	draw_projection(frame, R, landmarks, color)