Turning Natural Language into Art: DALL-E 2
What is DALL-E 2?
According to OpenAI, DALL-E 2 is:
"A new AI system that can create realistic images and art from a description in natural language."
Simply put, DALL-E 2 allows us to create original, realistic images and art from a simple text description.
What is Outpainting?
The DALL-E 2 outpainting tool allowed us to add context to a an image by providing a text description. We were curious to see how DALL-E 2 would contextualise real photographic images taken from the Perseverance Rover on the surface of Mars. Check out our gallery to see some of our best results.
How does DALL-E 2 work?
DALL-E 2 is a machine learning model that has learned the relationship between images and the text used to describe them. Specifically, the model uses a process called "diffusion", to progressively learn the patterns of colours and features within an image and its associated text description. DALL-E 2 uses over 3.5 billion parameters, trained on text-image pairs from the Internet.
How did we use DALL-E 2?
We provided a combination of different subject descriptions, using specific keywords and terms relating to space exploration, NASA and science. We also included various art styles, including Indigenous Australian Art, Abstract, Impressionist, Oil Paintings, and many more. All of our text descriptions are visible when you hover over the image in our gallery.
For more information about DALL-E 2 see the OpenAI Website:
For more information about the NASA Space Apps Challenge, see the website: