request: add detection level in image

[kalle07]

MediaPipe Solution (you are using)

mediapipe_face_mesh

Programming language

No response

Are you willing to contribute it

None

Describe the feature and the current behaviour/state

add detection level in image in a1111

Will this change the current API? How?

No response

Who will benefit with this feature?

No response

Please specify the use cases for this feature

for better conrtol

Any Other info

No response

[kuaashish]

Hi @kalle07,

We did not fully understand this feature. Could you please provide a brief explanation so that we can better understand it and bring it to our team's attention?

Thank you!!

[kalle07]

i see...

ok, the standart adetailer "face_yolov8n" shows a rectangle with the number of detection-level eg "0.71" so i can estimate which face has which level ;)

[kalle07]

mediapipe_face_full dont recognize upside down faces , why ?

[kuaashish]

Hi @kalle07,

Could you please let us know which solution you are requesting to add this feature to? Referring to the documentation, we believe you might be using the legacy Face Mesh solution. Could you please confirm this?

Thank you!!

[kalle07]

so this is a second point (besides detection-level)

its the standart in a1111 in adetailer

mediapipe_face_full

no option also face_yolov8n dont can handle upside down faces, it recognize but the image looks bad

[kuaashish]

Hi @kalle07,

Unfortunately, we cannot add this feature to the legacy face mesh solution you are using, as it is no longer maintained.

Thank you!!

[kalle07]

its mediapipe_face_full not mesh ...

btw, you work on somewhat new that will be implemented in adetailer ?

[kuaashish]

Hi @kalle07,

Thank you for the clarification. I am marking this feature and sharing it with the team. Based on future demand and discussions, we may implement it in the near future.