def transform_pose(self, pose):
pose_rot = quat2rotmat(pose[3:])
pose_trans = xyz2mat(pose[:3])
pose_cam = self.V.dot(pose_trans.T).dot(pose_rot).T
return np.concatenate(
[mat2xyz(pose_cam),
safemat2quat(pose_cam[:3, :3].T)])
def get_pose_in_camera(self):
mat = self.renderer.V.dot(self.pose_trans.T).dot(self.pose_rot).T
pose = np.concatenate([mat2xyz(mat), safemat2quat(mat[:3, :3].T)])
return pose
