Virtual Try-On Application

This is a virtual try-on application that allows users to upload their images and try on garments. The application integrates with Twilio for messaging and uses Gradio for virtual garment try-on. The process involves uploading a personal image and a garment image, followed by generating a prediction of the user wearing the garment.

System Diagram

Features

Twilio Integration: Receive images via WhatsApp and SMS.
Image Processing: Uses Gradio's Virtual Try-On model to generate predictions.
FastAPI Backend: Handles image uploads, processing, and responses.
Local File Hosting: Uses FastAPI's static file serving to host the generated images.
Prediction Generation: Uses OpenCV and Gradio to generate and store the final try-on image.

Installation

Clone the repository:
Install dependencies:
```
pip install -r requirements.txt
```

Create a .env file with the following environment variables:

TWILIO_ACCOUNT_SID=your_twilio_account_sid
TWILIO_AUTH_TOKEN=your_twilio_auth_token
NGROK_URL=your_ngrok_public_url

Set up Ngrok for tunneling to expose your local server to the internet:
```
ngrok http 8080
```

Start the application:

python3 -m uvicorn try_on.app:app --reload --port 8080

Usage

Send a WhatsApp message or SMS with an image of yourself (the person) to the Twilio number.
Once the person's image is received, the bot will ask for a garment image to try on.
After both images are uploaded, the application will process the images and send back a virtual try-on result with the user wearing the garment.

API Endpoints