Computers can now create art-like images with user input in mere minutes. We have been seeing AI-made art popping up quite often lately with Artbreeder or MidJourney, but DALL-E seems to be the best one yet since it is easy to edit what you prompt it to create.
DALL-E, which takes its name from Salvador Dali, the artist, and WALL-E from Pixar, is an artificial intelligence bot that can generate realistic and detailed images from simple or complex descriptions. You can ask for a rainbow-colored duck or an 18th-century steampunk mansion, and DALL-E can bring your imagination to life. In this article, we’ll take a closer look at DALL-E and what it is capable of with a couple of tricks to make it easier for you to come up with great prompts.
Generating images with DALL-E
For DALL-E to produce the most captivating images, you have to think about the way you give it the prompts. It helps to know some terms about art to get as close to the image in your mind. Since you only get 50 free credits to start with, the more specific and detailed you are with your description, the better.
Step 1: Describe the image in your mind
For example, let’s say I have a D&D game this weekend and I want my friends to have an idea of what my character looks like. If I am unable to draw or commission an artist, I can still get my idea on paper.
Start with your general idea and get more specific, such as:
“A futuristic cowboy riding an electronic horse synthwave” and so on. You can also state other styles such as “in the style of Van Gogh’s starry night, or pencil drawing”. Try to think of the lighting, the mood, and the style of the image.
Let’s see what that gives us.
Step 2: Fine tune your description
The second one is good enough, but this gave me another idea I would like to see so I will save the image and continue.
I used the prompt “A cowboy riding a horse in space synthwave” and ended up with these images:
Step 3: Add, remove, or change objects
I loved the second picture but I wanted a full moon on the upper left side, so I hovered over it and chose the “Edit Image” option.
I used the eraser tool to mark where I wanted the full moon to be and searched “full moon” on the search bar.
Step 4: Image generation is complete
And here is the final result:
Editing existing images
You can also use DALL-E to add creativity to existing images. You can upload a picture of yourself and change it up as you wish. You can be as realistic or as imaginative as you want.
Here is an example:
Do keep in mind that while using artificial intelligence for ideas like this is awesome, you should never use the images you create here commercially or portray them as your own painting to mislead people.