One cool thing we like about the Raspberry Pi is the sheer amount of resources available for integrating AI into your projects. Instead of using third party tools to relay AI tasks, you can actually host generative AI systems locally on the Raspberry Pi. Today's project is a wonderful example of how you can use this type of setup to create a beautiful, decorative display. Maker and developer Dylan has created what he calls PaperPi AI—a digital e-ink picture frame that creates flower paintings using AI.
Dylan is using a Pimoroni Inky Impression 7-color e-ink display as the main screen so the picture will stay even if the Pi loses power. The Pi generates pictures that look like paintings of flowers in various styles and updates the display with them periodically. It takes some time to generate the picture and update the screen so that does limit the number of changes, but it's not too bad for something that is chiefly intended as a decoration like this.
According to Dylan, it takes roughly half an hour to create a new image but you aren't limited to only generating flower paintings. You can program the Raspberry Pi to generate all sorts of topics and styles. Because the image generation takes place locally, you don't need an internet connection to make the project work. Once the Pi has created a new image, it takes 30 seconds to update the image on the e-ink display.
You don't need the latest, high-powered Raspberry Pi model to duplicate this project. Dylan is using a Raspberry Pi Zero 2 W to control the operation. It's connected to an Inky Impression 7.3-inch e-ink color display, which again supports seven separate colors. All of the hardware is housed inside of a picture frame big enough to mount the display panel as well as the Pi.
The Pi is running Raspberry Pi OS Bullseye Lite as the main operating system. The AI generation is handled using OnnxStream, this is a version of Stable Diffusion that's capable of running on the Raspberry Pi Zero. We covered this project back in 2023. You can read more about the project's development in greater detail and get a look at the source code over at the official GitHub project page.
If you want to get a closer look at this Raspberry Pi project, check out the official thread shared to Reddit for more information.
