xijunjun / dlib-face-recognition-delaunay-triangulation

Python project to perform face detection and landmarks recognition using dlib plus delaunay triangulation using OpenCV

Dlib Face Detection + OpenCV Delaunay Triangulation

This project aims to guide you in the process of face detection and landmark recognition, using Dlib library. In addition, it shows how to visualize the results using OpenCV and using two techniques known as: delaunay triangulation and voronoi diagrams.

See article on Medium for a more detailed explanation.

Requirements

Before running anything, you must install the project dependencies:

opencv-python==4.5.1.48
numpy==1.19.5
dlib==19.21.1
imutils==0.5.4

You can run pip install to install all the dependencies at once or install it one by one. This process may take a few minutes.

pip install -r requirements.txt

Usage

The project has two main modules: face_recognition.py where you will find everything related to face detection and landmark recognition, and delaunay.py where you will find the main script, delaunay triangulation, etc.

To run face detection and analysis for example for Luis.jpg photo, call delaunay.py as follows:

python delaunay.py --image Luis.jpg 

Also, you can change script behavior using different flags:

• --image : Path to image to process
• --vovonoi : Display face landmarks using Voronoi diagram
• --save : Save results as images
• --l28 : Use only 28 of the 68 face landmarks returned by Dlib shape_predictor.