A headline without an appropriate accompanying image isn't a good way to digest your news. But what do you do if the article doesn't have an image? Generate one using a Raspberry Pi with AI, of course! Maker and developer Roy van der Veen has created a project that presents the generated image alongside the headline text that led to its creation on an E-ink display.
According to Van der Veen, the project relies on Stable Diffusion. Headlines are selected randomly every 5 hours. Stable Diffusion uses this text to create an image based on its interpretation. The idea was to create something fun and somewhat artistic rather than a platform for actually reading the news. As a result, the images are displayed in an E-ink frame, rotating throughout the day.
It's also worth noting that the AI-generated images aren't saved on the Pi. The only way to experience the picture is to be there when it's created. However, you could modify it to save the images so you can pick out your favorites. Van der Veen also points out how fun it can be to notice the little discrepancies in what the AI interprets from the headline text.
The hardware is housed inside a modified IKEA picture frame with a Raspberry Pi Zero 2 W mounted to the outside. This keeps the Pi relatively cool, but you could always use a fan or some other cooling mechanism and place the Pi inside. The images are output to a 7.3-inch Inky Impression frame from Pimoroni.
The source code for the project has been made open source. You can find details about the project and get a closer look at how it works at GitHub. Python appears to be the language of choice for this project, controlling the output on the eInk display. Instructions on installing the software are also available for anyone who wants to recreate the project.
You can read more about this Raspberry Pi project at GitHub. There are plenty of hilarious examples of what this system is like in action. You can also follow Roy van der Veen for future creations and any updates to this one.
