Story Image Generator

Motivation

Our group understands the pains of creating eye-catching illustrations to accompany literary masterpieces. The process of transforming words into compelling visuals is both time-consuming and often requires a high level of artistic skill. For authors and content creators, this can be a significant barrier, potentially limiting the impact and reach of their work.

We believe that every story deserves to be brought to life with stunning visuals that captivate the audience and enhance the narrative experience. Our project aims to simplify the creation of illustrations, enabling writers to effortlessly generate beautiful images that complement and elevate their literary creations.

By leveraging advanced AI technologies, we provide an innovative solution that bridges the gap between text and imagery. Our tool empowers authors to focus on their storytelling while seamlessly generating illustrations that resonate with their readers, making the process of creating illustrated literary works more accessible and enjoyable.

Functionality

This application generates images based on a user-provided story by leveraging OpenAI's GPT model and Bing Image Creator accessed via a reversed engineered API.

• Story to Prompts: The OpenAI GPT model is used to analyse the user-provided story and generate concise, descriptive prompts that are then used for image generation. These prompts capture key scenes and elements of the story to ensure that the generated images accurately reflect the narrative.

• Image Generation: The generated prompts are then passed to the Bing Image Creator via its API. High-quality images based on the provided prompts.

• Displaying Results: The resulting images are displayed, providing visually compelling illustrations that complement and enhance the original story. The authors can then use the imagesto complement their works.

Setup

Prerequisites

• Python 3.6 or higher
• Google Chrome browser
• Git
• Azure API Key from OpenAI

Installation

1. Clone the repository:

   Clone the repo and follow the instructions below to set up.

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. Create a .env file in the root directory of the project and add your API keys and other necessary environment variables. You can use the .env.example file as a reference:

   cp .env.example .env

5. Open the .env file and add your API keys and other environment variables:

   AZURE_API_KEY='replace_with_your_azure_api_key'
   AZURE_ENDPOINT='replace_with_your_azure_endpoint'

Running the Application

1. Run the Flask application:

   python app.py

2. Open your browser (Google Chrome recommended) and navigate to http://localhost:5000.

3. Follow the instructions on the webpage to generate images based on your story.

Project Structure

• app.py: The main Flask application file.
• authcheck.py: Handles Bing login and cookie extraction.
• BingImageCreator.py: Contains the Bing Image Creator integration.
• llm.py: Handles prompt generation using OpenAI's GPT model.
• templates/: Contains the HTML templates.
  • home.html: The homepage where users input the number of images and their story.
  • instructions.html: Instructions for logging into Bing.
  • display_images.html: Displays the generated images.
• .env.example: Example environment file to help users set up their own .env.

Acknowledgements

This project makes use of the following third-party library:

• Bing Image Creator - The use of their API to access Bing's AI image generator.
• OpenAI - The use in code for generating text prompts based on the story input.