U.S. Cash Currency Detection

By simplifying currency identification, this project promotes financial accessibility, inclusion, and independence—all of which align with the core goals of social good.

Comment

Inspiration

To make currency recognition accessible and easy, especially for those with visual impairments, language barriers, or limited financial literacy.

What it does

Detects the denomination and quantity of U.S. cash from PNG, JPG, or JPEG images and displays predictions with confidence percentages.

How I built it

I created a dataset of 1,000+ images with Roboflow, trained a machine learning model, built an HTML frontend, connected it to a Flask backend, and deployed via Onrender.

The model is accessible here: link

Once model was complete, I utilized HTML to create a simple frontend. It consisted just of a page that outlined my project and allowed for file uploads. For the backend I used Flask to temporarily save the image and send it my own Roboflow model. Lastly, I pushed this all to GitHub and deployed it with Onrender to get my finished product.

How accurate is this model?

My model has a mAP@50 of 94.0% Where mAP@50 is equal to the mean of the Average Precision metric across all classes in a model at a 50% IoU threshold. And has a precision of 90.7%. Which measures how often a model's predictions are correct. Lastly, it has a recall of 88.0%. Where recall measures what percentage of relevant labels were successfully identified.

For the Youth Coders Hack 2025, how does this relate to the topic of "Social Good"?

This project contributes to Social Good by making currency recognition more accessible and efficient. By allowing users to upload an image and instantly identify both the denomination and quantity of U.S. currency, the tool can support individuals with visual impairments, cash handling challenges, or language barriers.

It can be used by educational programs, nonprofits, or financial literacy initiatives to help people better understand and manage physical money. The website is fully online, easy to use, and integrates Roboflow for accurate predictions.

By simplifying currency identification, this project promotes financial accessibility, inclusion, and independence—all of which align with the core goals of social good.

Challenges I ran into

I was a beginner to HTML, Flask, and Python, so learning how to connect the frontend, backend, model, and deploy online was challenging.

Accomplishments that I'm proud of

I managed to successfully build and deploy a fully functional AI-powered currency detector accessible online.

What I learned

How to integrate a ML model with Flask, handle file uploads, deploy a web app, and create a user-friendly frontend.

What's next for U.S. Cash Currency Detection

Add coin recognition, support other currencies, improve UI accessibility, add text-to-speech for visually impaired users, and expand the dataset for higher accuracy.

Built With

Share this project:

Updates