← Back

ofxMediaPipePython

github.com/design-io/ofxMediaPipePython · @design-io
OF 0.12 Machine Learning Computer Vision

Addon for running python version of MediaPipe within an OpenFrameworks application using pybind.

10 1 Last commit 7mo ago Created 1y ago

README

ofxMediaPipePython

Addon for running python version of MediaPipe within an OpenFrameworks application using pybind.

ofxMediaPipe Example

Tracking

Faces: https://ai.google.dev/edge/mediapipe/solutions/vision/face_landmarker/python

Hands: https://ai.google.dev/edge/mediapipe/solutions/vision/hand_landmarker/python

Poses: https://ai.google.dev/edge/mediapipe/solutions/vision/pose_landmarker/python



Platforms

Supported on macOS and Linux using OpenFrameworks >= 12.0.

An install script has been provided to help with installation of Media Pipe. The script will download miniconda, create a conda environment named 'mediapipe' and install mediapipe using pip.

To install:

cd to ofxMediaPipePython

./InstallMediaPipe.sh

Build the example projects using the Project Generator.

ofxMediaPipe Example

OSC Sending and Receiving

ofxMediaPipe OSC TX RX Example

Compiling and running the OSC sender and receiver examples require the Preprocessor macro OF_ADDON_HAS_OFX_OSC

XCode:

Navigate to the Xcode project settings / Build Settings / Preprocessor Macros and add OF_ADDON_HAS_OFX_OSC=1 Preprocessor macro in XCode

Linux:

Add OF_ADDON_HAS_OFX_OSC to the project config file, for example: MediaPipeOscReceiverExample / config.make

Adjust it to the following ( removing the # sign )

PROJECT_DEFINES = OF_ADDON_HAS_OFX_OSC

Get this addon

HTTPS
SSH