Photohm

Android application that allows you to detect components of an electrical circuit from a photograph or screenshot, build the corresponding graph and simulate its electrical behavior.

Comment

Inspiration

As students, we saw a real problem with circuit simulators: the amount of time it takes to build a circuit from scratch just to test it and see the results. On top of that, existing AI tools struggle to accurately detect and interpret photos of real circuits, making the process even slower and more frustrating.

What it does

Photohm lets you take a photo of a physical circuit and, using Gemini, accurately identifies and explains the circuit's components and connections. From there, you can make modifications and test changes without having to rebuild the circuit manually in a traditional simulator.

How we built it

We combined computer vision for component detection with Gemini for circuit interpretation and explanation, allowing the system to go from a raw photo to an editable, testable circuit representation.

Challenges we ran into

We initially used YOLO for component detection, but had to switch to YOLOX due to licensing restrictions.

Accomplishments that we're proud of

Getting the model to reliably detect not just clean screenshots of circuits, but real, imperfect photos taken by users.

What we learned

We learned how to work with AI models hands-on and push them to their full potential for a real-world use case.

What's next for Photohm

Bring Photohm to market. Add support for new components since the core logic is already built, expanding functionality will be faster and easier. Submit to other platforms

Share this project:

Updates