AI Camera
An AI camera that regenerates the photos you take to fit a theme.
This project began as a creative concept that I decided to bring to life. The camera captures a photo, generates a text description of it using AI, then produces a new image based on that description through a second AI. Users can select a visual theme before taking each shot.
All structural parts are 3D-printed and painted. The system runs on a Raspberry Pi Zero with a camera module, connects to AI services via API, and uses a phone hotspot for internet access.
The camera pipeline runs two AI models in sequence: the first converts the captured photo into a text description, and the second generates a new image from that description with the user-selected theme applied. The result is an image that shares the composition of the original but is visually reimagined.
A GUI running on the camera's display lets the user select a theme before shooting. The interface was designed to be simple enough to operate with physical buttons while providing meaningful control over the generation style.
The system runs on a Raspberry Pi Zero with a custom PCB handling power management and peripherals. Battery access is through a panel on the bottom of the enclosure. All structural components are 3D-printed and painted.
