I started experimenting with a new artificial intelligence art app the other day called IMGCreator, and at that time I was using their feature that allows one to use an existing picture and have their app generate a new image based on part of the existing one. I was taking a Van Gogh self portrait and letting IMGCreator come up with new images extending off of Vincent Van Gogh’s face. But today, I wanted to try creating artificial intelligence art directly from text keywords.
I decided to see what sort of realistic-looking objects I could create from a variety of keywords and phrases. The idea is that the artificial intelligence is using algorithms to pull from a database of imagery to create something totally new that doesn’t visually exist (at least not exactly as depicted in the end result). Here’s four initial results from today’s experiments with IMGCreator, clockwise from top left, the keywords were: Aztec figure; Murano glass sculpture on a pedestal; gothic futurist mask on a pedestal in a museum with red walls; and elaborate wood sculpture on a pedestal.
My process for making these went like this: first, I entered my keywords or phrase; then I get to choose from a number of categories, such as 3-D render design, realistic photo, illustration (and more); then I can choose a style, such as medium shot, close-up, low angle, etc.; and lastly I can set a “canvas size,” which can be square, vertical portrait or horizontal landscape mode. It’s all very easy to use, and it only takes a few seconds for the images to generate, and you get two examples at a time.
So, what do you think? Are they convincing? Share your thoughts in the comments section below. If you’re interested in trying this, check out their options here.