Raspberry Pi AI EYE camera regenerates pictures using an AI description

Oscar Wilmerding’s AI EYE is an innovative Raspberry Pi-based camera project that integrates AI to transform captured images into entirely new creations. This project combines the functionality of a standard camera with AI-based image processing to provide a unique twist: after snapping a picture, the AI scans the image, generates a detailed textual description, and then uses that description to create an entirely new image, sometimes with an added thematic flair based on user input.

For instance, if you photograph a cat, the AI can recreate it with a “spooky Halloween” theme. The project relies on two main AI services: Astica Vision for generating image descriptions and DALLE 3 to transform the text into a new image.

The camera is built using a Raspberry Pi Zero 2 W, chosen for its portability, though it delegates the heavier AI tasks to online services due to the Pi’s limited computing power. Housed in a custom 3D-printed shell, the camera includes a 3.5-inch LCD for real-time output and is powered by a 3.7V lithium-ion battery. It also features a custom PCB to handle the power and input connections.

Wilmerding has made the project open source, providing all the necessary details on GitHub and Reddit for DIY enthusiasts to recreate it at home. He also hinted at the possibility of releasing kits for easier assembly in the future, depending on interest.

This project is a great example of using AI creatively with Raspberry Pi, offering makers an exciting way to explore AI-driven art and photography.

Read more: Raspberry Pi AI EYE camera regenerates pictures using an AI description