Simplified Text-to-Image Generator A beginner-friendly Python project that generates simple images based on text prompts, showcasing the basic principles of text-to-image systems.
This project demonstrates how to convert a text input into an image by:
Mapping Text to Numerical Embeddings: Converting the input text into a numerical representation. Mapping Embeddings to Visual Parameters: Using the embedding to define attributes like shape, color, and size. Generating Images: Drawing simple shapes (circles or squares) based on the extracted parameters.
The goal is to provide an educational starting point for understanding text-to-image generation.
Generates images with one shape: circle or square. The shape's color and size are dynamically derived from the input text. Supports customizable canvas size.
Clone the repository:
git clone https://github.com/sarkarbikram90/imagen.git
Navigate to the project directory:
cd imagen
Run the script:
python imagen.py
Enter a text prompt when prompted. The generated image will be displayed in a new window.
The image generation is based on simple heuristics, not advanced machine learning. Only basic shapes (circle or square) are supported.