Neural Style Transfer with Flask

This project demonstrates a neural style transfer application using TensorFlow and Flask. The application allows users to upload content and style images and produces a stylized image that combines the content of the content image with the style of the style image.

Prerequisites

• Python 3.6 or higher
• TensorFlow
• TensorFlow Hub
• Flask
• Flask-WTF
• Matplotlib
• Numpy

Installation

1. Clone the Repository

   git clone https://github.com/koushik2k3/Neural-Style-Transfer.git
   cd Neural-Style-Transfer

2. Create and Activate a Virtual Environment

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

3. Install the Required Packages

   pip install tensorflow tensorflow-hub flask flask-wtf matplotlib numpy

Usage

1. Start the Flask Application

   python app.py

   The Flask app will start on http://127.0.0.1:3000/.

2. Upload Images

   • Navigate to the root URL of the Flask app.
   • Use the form to upload a content image and a style image.
   • Submit the form to process the images.

3. View the Result

   • The processed stylized image will be displayed on the page.
   • The stylized image will also be saved to the static/output directory.

Project Structure

• app.py: Main Flask application script.
• templates/: Contains the HTML templates.
• static/: Contains static files including uploaded images and output images.