Constant loss against bluffers while playing poker.

What it does

This app simply focuses on detecting the emotional features of a person and decides if the person is bluffing or not during the game!

How I built it

We used Emotion API by Microsoft with a small addition of PHP / HTML / CSS / JAVASCRIPT for representing the graphical interface of the AI system, where you simply upload a picture of a person (who might be bluffing or not) & system would simply tell you if the person is a bluffer or not.

Challenges I ran into

Microsoft API was not compatible with Python while working using a Mac computer, in addition, the major challenge we've had was collecting the right images to train our AI system to have some decision making abilities.

Accomplishments that I'm proud of

System that we've built was quite reliable, in a sense that, it would identify (85% of the time) if the person is a bluffer or not, so you could use this app to either improve your ability to spot people who bluff during the game or simply have fun watching the system tackle emotions of professional poker players.

What I learned

We've learnt a lot about implementing Microsoft Cognitive Services into our projects, also, using REST APIs, processing JSON formats, sending HTTP requests / getting some responses. We've also improved our attention to detail, enhanced our PHP development knowledge, team working skills.

What's next for PokerFace

Setting up a server to be able to hold this system remotely, ability for this website to be able to access it from a domain. One of the most important improvements would be to feed much more data into the AI system, so our predictions would be more precise and less error-prone! Hopefully, this will help future players decide if their opponent is lying!

Built With

Share this project: