Janus AI - Face Swapping using Mediapipe and OpenCV

Janus AI is a Django-based application leveraging the capabilities of Mediapipe and OpenCV to facilitate face swapping in both images and video.

Overview

This project demonstrates how to perform face swapping in real-time videos using facial landmarks. It uses the mediapipe library to detect facial landmarks and OpenCV for image processing.

Getting Started

Clone the repository:

git clone https://github.com/GBU31/Janus_AI.git

Install the required dependencies:

pip install -r requirements.txt

Apply migrations:

python3 manage.py migrate

Run the Django server:

python3 manage.py runserver

API Usage

Example using cURL

curl -X POST -H "Content-Type: multipart/form-data" -F "image_one=@./image_one.jpg" -F "image_two=@./image_two.jpg" http://localhost:8000/api

Example using Python requests

import requests

url = "http://localhost:8000/api"
files = {
    "image_one": ("image_one.jpg", open("image_one.jpg", "rb")),
    "image_two": ("image_two.jpg", open("image_two.jpg", "rb")),
}

response = requests.post(url, files=files)
print(response.json())