Vi-Tons Imagen

Vi-Tons Imagen is an AI-powered virtual try-on application that allows users to generate realistic human models or edit reference images with customized clothing and accessories. The application leverages Imagen 3 model and integrates with Google Cloud Storage for seamless image processing and storage.

Features

• Generate Human Models: Create realistic human models based on user-provided attributes like gender, ethnicity, age, body type, and clothing preferences.
• Reference Image Editing: Upload a reference image to apply custom clothing and accessories.
• Dynamic Prompt Refinement: Refine user inputs into AI-ready prompts for generating or editing images.
• Google Cloud Storage Integration: Store and retrieve generated images securely using Google Cloud Storage.
• Image Download: Serve generated images directly for download.

Code Repository Structure

The repository is organized as follows:

vi-tons_imagen/
├── app.py                     # Main FastAPI application file
├── README.md                  # Project documentation
├── requirements.txt           # Python dependencies
├── app.yaml                   # Configuration file for Google App Engine deployment
├── .env                       # Environment variables (not included in the repo)
├── prompts/                   # Directory for prompt templates
│   └── prompt_templates.py    # Predefined templates for AI prompts
├── utils/                     # Utility functions
│   ├── helper_functions.py    # Helper functions for AI model interactions and image processing
│   ├── constants.py           # Constants and configuration variables

Setup

Prerequisites

• Python 3.8 or higher
• Google Cloud account with access to Google Cloud Storage
• Google Cloud SDK installed
• A Google Cloud Storage bucket for storing generated images

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/vi-tons_imagen.git
   cd vi-tons_imagen

2. Create a virtual environment and activate it:

   python -m venv venv
   source venv/bin/activate  # On Windows, use `venv\Scripts\activate`

3. Install the required dependencies:

   pip install -r requirements.txt

4. Set up environment variables:

   Create a .env file in the root directory and add the following variables:

   GOOGLE_PROJECT_ID=your-google-project-id
   GCP_BUCKET_NAME=your-google-cloud-storage-bucket-name
   

5. Authenticate with Google Cloud:

   gcloud auth application-default login

Usage

Running the Application Locally

1. Start the FastAPI server:

   uvicorn app:app --reload

2. Open your browser and navigate to http://127.0.0.1:8000/docs to access the Swagger UI and test the API endpoints.

API Endpoints

POST /generate_tryon

Generate a try-on image based on user input or a reference image.

• Request Parameters:

  • user_request (Form): JSON string containing user input attributes.
  • reference_image (File, optional): Reference image with a full-body pose.

• Response:

  • If a reference image is provided: Returns the URL of the edited image.
  • If no reference image is provided: Returns the URL of the generated human model.

GET /download_image

Download a generated image from Google Cloud Storage.

• Request Parameters:

  • image_url (Query): The GCS URL of the image to download.

• Response:

  • Returns the image file for download.

Example Requests

Generate Try-On Image (Without Reference Image)

curl -X 'POST' \
  'http://127.0.0.1:8000/generate_tryon' \
  -H 'accept: application/json' \
  -H 'Content-Type: multipart/form-data' \
  -F 'user_request={
        "user_input": {
            "gender": "male",
            "ethnicity": "Indian",
            "age": "30",
            "body_type": "athletic",
            "mood": "confident",
            "style": "casual",
            "top": {"color": "yellow", "type": "t-shirt"},
            "bottom": {"color": "black", "type": "jeans"},
            "footwear": {"color": "yellow", "type": "sneakers"},
            "accessories": {"colour": black", "type": "watch"}
        }
    }'

Generate Try-On Image (With Reference Image)

curl -X 'POST' \
  'http://127.0.0.1:8000/generate_tryon' \
  -H 'accept: application/json' \
  -H 'Content-Type: multipart/form-data' \
  -F 'user_request={
        "user_input": {
            "style": "formal",
            "top": {"color": "black", "type": "blazer"},
            "bottom": {"color": "black", "type": "trousers"}
        }
    }' \
  -F 'reference_image=@path_to_your_image.jpg'

Download Image

curl -X 'GET' \
  'http://127.0.0.1:8000/download_image?image_url=gs://your-bucket-name/path-to-image.jpg' \
  -H 'accept: application/json'

Deployment

Deploying to Google App Engine

1. Create an app.yaml file in the root directory with the following content:

   runtime: python39
   entrypoint: uvicorn app:app --host 0.0.0.0 --port $PORT
   
   env_variables:
     GOOGLE_PROJECT_ID: your-google-project-id
     GCP_BUCKET_NAME: your-google-cloud-storage-bucket-name
   
   handlers:
   - url: /.*
     script: auto

2. Deploy the application:

   gcloud app deploy

3. Access the deployed application at the URL provided by Google App Engine.

Contributing

Contributions are welcome! Please open an issue or submit a pull request for any improvements or bug fixes.

License

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