I enjoy tech and development, I also enjoy playing poker occasionally. So when I find areas where both those things overlap, it interests me.
Microsoft have been developing a number of what they call cognitives services which allow the user to take advantage of some clever machine learning technology Microsoft have built.
One of these services is called the Emotion API which provides the ability to identify faces in an image and calculate which emotions the person is showing on their face.
Each face displays a certain amount of the following emotion.
So I thought it would be fun to run this tool against some poker players poker faces.
For those people unfamiliar with the term, a poker face is the face a poker player shows when playing poker.
A perfect poker face would not give anything away to your opponent, there would be no sign of pleasure or dissapointment in with the cards they have or the hand they are in. The aim is to look completely neutral.
So i'd expect to see a high level of neutrality in the poker players faces according to the Emotion API.
Let's see the results..
92% neutral, 4% contempt, 1% anger
94% neutral, 3% contempt, 2% happiness
As you can see, a high level of neutrality in all 3 poker players.
Interesting stuff, I may do a follow up blog post comparing dozens of poker players poker faces.. If I have time.