Having worked quite some time remotely I know of the annoying practice of keeping every receipt in order to get your expenses back. Further we see the potential to establish a loyalty program based on receipts for small to medium enterprises. That's how we had the idea to build an ML based receipt scanner that seamlessly sums up your expenses.

What it does

Films receipt, finds text & numbers, sums up expenses, saves picture of given bill with matching information. Saves time.

How I built it

Thanks to our friends at Huawei we were able to test their HMS Core ecosystem on a device generously given to us for the duration of HackZurich. By using the ML Kit framework which lives in HMS Core we built a receipt scanner with ease in a language previously unknown to us, Kotlin.

Challenges I ran into

Working with diverse previously unknown frameworks, languages and paradigms.

Accomplishments that I'm proud of

Working with diverse previously unknown frameworks, languages and paradigms.

But seriously, finishing an MVP on time in an area (ML / Computer Vision) we both were not acquainted with.

What I learned

Kotlin, Android Studio, HMS Core -- ML Kit and some Java.

What's next for Burning Expenses

We see a strong use case for our solution to the expense summation problem therefore we will be looking into further optimization and development of our application.

PS We don't have video editing software which is why we put it up our pitch and the demo together in a playlist -- we hope you don't mind!

Built With

Share this project: