cross-platform, customizable ML solutions for live and streaming media. - Home / github / Introduction to MediaPipe

MediaPipe uses TensorFlow lite in the backend.

see also

• Human Pose Estimation using OpenCV & Python
• Gesture Volume Control Using OpenCV and MediaPipe

GPU Support

MediaPipe supports OpenGL ES up to version 3.2.
MediaPipe framework doesn’t require CUDA for GPU compute and rendering. However, MediaPipe can work with TensorFlow to perform GPU inference on video cards that support CUDA.

MediaPipe in C++

MediaPipe Face Mesh

Installation

using Docker