why iris tracking do not take account of the rotation of head

[dandingol03]

OS Platform and Distribution

OSX

Task name (e.g. Image classification, Gesture recognition etc.)

iris tracking

Programming Language and version (e.g. C++, Python, Java)

c++

Describe the actual behavior

In the process of iris-to-depth-tracking, i saw the computing of the iris diameter only consider the x span and y span between iris top landmark and iris bottom landmark. If the face is not exactly facing the screen or there is some rotation of the head in the x-z axis, the iris size would be inaccurate.

inline float GetLandmarkDepth(const NormalizedLandmark& ld0,
                              const NormalizedLandmark& ld1,
                              const std::pair<int, int>& image_size) {
  return GetDepth(ld0.x() * image_size.first, ld0.y() * image_size.second,
                  ld1.x() * image_size.first, ld1.y() * image_size.second);
}

Describe the expected behaviour

Take account of the head pos in iris tracking. For example, compute the projected size of iris in x-y plane, and project back to the corret size with face rotation.

[kuaashish]

Hello @dandingol03,

We are upgrading the MediaPipe Legacy Solutions to new MediaPipe solutions And Now, Iris tracking is part of the Face landmark detection as guided here.

These new solutions will provide a superset of capabilities available in the legacy solutions. We recommend you to explore the our new upgraded solution and revert back if you observe the similar behaviour. Thank you!!

[github-actions[bot]]

This issue has been marked stale because it has no recent activity since 7 days. It will be closed if no further activity occurs. Thank you.

[github-actions[bot]]

This issue was closed due to lack of activity after being marked stale for past 7 days.

[google-ml-butler[bot]]

Are you satisfied with the resolution of your issue? Yes No