Artist Uses AI to Extract Color Palettes from Text Descriptions
Reading Time: 2 minutes

Matt DesLauriers, a London based artist uses AI to extract color palettes from text description. The tool he developed allows anyone to type in “beautiful sunset” and get a series of colors that match a typical sunset scene. The tool can also find colors that match a sad and rainy Tuesday.

DesLauriers has posted his code on GitHub. It requires Stable Diffusion, Node.JS, and a JavaScript GIF encoder named gifenc extracts the palette information. The images generated are impressive examples of open source image synthesis models.

In a tweet DesLauriers demonstrated how different quantization methods could produce different color palettes.

Introducing DesLauriers’s different palettes. For Tokyo, he suggests colors from a vibrant cityscape with deep pinks and blue hues, while for a coral reef there are deep pinks and blues that represent the scene. DesLauriers has also presented how different color quantization methods can produce slightly different palettes

Some other artists have also tried AI-generated color palettes from text. For example, dribnet published a generative art series called “Homage to the Pixel.” Through this, anyone can make a six-color palette by inputting text.

Why should you use an AI to find color palettes? Aside from novelty, a machine can find colors that match styles and philosophies. One example is when a machine selects a color palette based on the idea of “the day after my last day in high school”.

Image synthesis models are likely to be used in various applications that are new, such as color selection. With the help of an algorithm, you can get what you’re looking for.

Related Articles:
New AI-Powered X-Ray Technique Helps Detect Explosives and also Identify Cancer
Electric School Buses are Cheaper and Easy to Maintain with Couple of Drawbacks
KEBOT – An Advanced Robotic Scanning System For Hair Transplantation