Typescript/Javascript implementation of Face Clustering using ONNXRuntimeWeb

Description

This project is an implementation of the Multi-task Cascaded Convolutional Networks (MTCNN) for face detection and alignment and VGGFace for calculating embeddings.

Based on MTCNN Python implementation

Based on https://github.com/yiyuezhuo/mtcnn-onnxruntime

• Python=3.7
• pip install onnxruntime opencv-python
• python3 app.py

Files

• mtcnnjs/pnet.js: Contains the JavaScript implementation of the P-Net. (Stage 1)
• mtcnnjs/rnet.js: Contains the JavaScript implementation of the R-Net. (Stage 2)
• mtcnnjs/onet.js: Contains the JavaScript implementation of the O-Net. (Stage 3)
• mtcnn_ort/mtcnn_ort.py: Contains the Python implementation of MTCNN using ONNX Runtime.

Usage

To use this project, you will need to install the necessary dependencies, then run the appropriate script for your use case. You need to run Google Chrome/Canary with chrome://flags/#enable-webgpu-developer-features enabled.

• npm install http-server
• http-server . -p 8080
• navigate to http://localhost:8080/app.html