For face landmarker, please provide info on optimal input image size, face size, face location in image, etc.

[by321]

Description of issue (what needs changing)

missing information

Clear description

Face landmarker seems very sensitive to size of input image, face, or face locations within the image.

Correct links

No response

Parameters defined

No response

Returns defined

No response

Raises listed and defined

No response

Usage example

No response

Request visuals, if applicable

No response

Submit a pull request?

No

[kuaashish]

@by321,

Thanks for raising this request. We are forwarding issue internally and based on discussion team will prioritise the request.

[kuaashish]

Hello @by321,

The current face detection model is using the short range model, https://developers.google.com/mediapipe/solutions/vision/face_detector#blazeface_short-range.

The model is trained to used for detecting human face in a short range, more specifically from the model card: https://storage.googleapis.com/mediapipe-assets/MediaPipe%20BlazeFace%20Model%20Card%20(Short%20Range).pdf

In out-of-scope setion: ● Detecting people too far away from the camera (e.g. further than 2 meters)

There should be no limitation regarding to the image size and the face location in the image. The model input size is 128x128, and the pipeline will resize the input image to 128x128. 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.