Aces Over

• 

  In this case, you should hit

Inspiration

Our team has always been interested in learning more about game theory and how the casinos cheat you.

What it does

For this reason, we wanted to create an app that could help educate you in card counting in blackjack to turn the odds in your favor.

How we built it

We built this app using React Native with Expo for the frontend/app interface. We used Flask for our backend api, and used a custom pre-trained model for classifying playing cards. We sourced this model from the open Playing-Cards-Object-Detection repository on GitHub.

Challenges we ran into

During our building of this app, we ran into multiple challenges. Due to our lack of experience with mobile development, we struggled a lot with getting accustomed to our frontend toolset. We also had a lot of issues finding proper documentation for the camera module that we made a lot of use of. Additionally, it was difficult to get our model to consistently recognize cards. We initially planned on using an online api, but it wasn't consistent enough, so we decided to run a local model. We also had to take bursts of photos rather than individual photos to improve the consistency of the classifier.

Accomplishments that we're proud of

We were particularly proud of our quick ability to pivot to a different model, as well as to improve our apps performance by implementing burst photos.

What we learned

We learned a ton about mobile development, as well as about hosting and integrating with classification models.

What's next for Aces Over

Our next plan would be to improve the camera so that it recognizes moves in real time, without the requirement for user input. We would also like to increase the accuracy of our playing card classifier.

Built With

• expo.io
• flask
• playing-cards-object-detection
• react-native