MediaPipe API Server

Get Started

pip install -r requirements.txt -f https://mirror.sjtu.edu.cn/pytorch-wheels/torch_stable.html

Acquire ckpt for mmpose via https://github.com/open-mmlab/mmpose/tree/main/projects/rtmo and put it in ./ckpt. After that, select the corresponding configuration file, the default configuration is for "rtmo-t_8xb32-600e_body7-416x416-f48f75cb_20231219.pth"

Start server

python -m mediapipe_apiserver serve

Start Client

wscat -c 127.0.0.1:3000/v1/dummy
python ./playground/test_ws_client.py

TODO

• [√] update requirements
• raw image api, annotated image API
• client library
• add __enter__ to camera class that supports with context

Develope Guide

Requirements

• A Azure Kinect camera.
• Azure Kinect SDK Installed. See https://learn.microsoft.com/en-us/azure/kinect-dk/sensor-sdk-download.
• A Websocket debug tool, such as wscat(Run npm -g install wscat to install).

Licenses

This project is licensed under the MIT License - see the LICENSE file for details.

This project uses MediaPipe, which is licensed under the Apache License 2.0.