Skip to content

Is it possible to manually map face landmarks for images where mediapipe doesn't detect them at all? #5519

Open
@aesanchezgh

Description

@aesanchezgh

Hi all,

I have some difficult images where mediapipe is not detecting the face landmarks. If I lower the thresholds and increase the image brightness, it detects landmarks, but they are not accurate. Is there any method where I could use python code in some way to add all 478 points to an existing image I have?

Thanks!

Metadata

Metadata

Assignees

Labels

task:face landmarkerIssues related to Face Landmarker: Identify facial features for visual effects and avatars.type:supportGeneral questions

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions