-
### Functionality details
1. add a api endpoint which is able to detect multiple human faces in the image
2. May use tensorflow ml models for the task or import other 3rd party models
3. **Optio…
-
First, I express my gratitude for the dedication and hard work of all developers!
During my self-study time in Mediapipe, I have been familiar with the most of examples from "mediapipe/examples/des…
-
Online demo at https://huggingface.co/spaces/jbilcke-hf/FacePoke does not work on this face for example.
It shows orange face outline but grabbing and moving a mouse cursor around does nothing.
!…
-
might be interesting to use something like https://google.github.io/mediapipe/ and track movement of objects with images… for the brain theres (probably??) nothing moving as much for smtg like this to…
-
I'm writing to report an issue I've encountered while working with the tasks-vision.facelandmarker in MediaPipe. I am trying to overlay virtual glasses on a user's face, and while everything works per…
-
This is to try out the gesture recognition for the user input.
-
**Question :**
I am impressed with your project and find it very interesting.
However, after pushing the model(scigemma.bin) and build to Android Studio(modify InferenceModel.kt), upon running the …
-
Hi, this seems to work pretty well! I'm sending the OSC message to Wekinator, but it seems that the incoming OSC messages misses some values...
1. it seems it doesnt count the hands ((i always get …
-
I build the application according to the guide, but the resolution is so low.
by the way, my phone is xiaomi10, and the resolution of my phone is 1920*1080p.
Any solution? pls!
-
### Have I written custom code (as opposed to using a stock example script provided in MediaPipe)
None
### OS Platform and Distribution
Linux
### Python Version
3.10
### MediaPipe Mo…