We started off with a brainstorming session which focussed on the idea of limited information games, and ways to add more information to them. This included quantum-based variations on classics like chess, probabilistic connect four and a facially controlled version of a poker variant. After exploring these ideas we realised that quantum computing wasn't practical for the hack and so focussed on facial recognition. This lead us into the restrictions a face-based input system would bring, such as people being unable to use it, which lead us to develop the idea of its usage in emotional development, although that was never completed.
What it does
Using Azure's cloud computing we convert live video (from HTML5) into a set of parameters for manipulating game inputs (in python), allowing you to play these games with facial expression alone, no head movement necessary. It currently only runs two
Challenges I ran into
Integrating both the Flask library for Python and the Azure cloud led to some complex data parsing problems between multiple different environments. It took a significant amount of time to figure out how to use all 4 engines in real time in parallel on one machine. Group coordination and division of labour was also an issue due to the scale of apparently simple tasks becoming much larger than expected.
Accomplishments that I'm proud of
It works. It's not finished, and it appears mediocre at best, but the backend technology is there, and it is functional and of that I am proud.
What I learned
As a team we certainly could have achieved more if we had a clear idea of how much work each part of the construction of our program would take, and planned for a slightly bigger project.
What's next for Face InterFace
Analytics. Face InterFace currently produces data for general use, but without any field experience the data has nothing to compare to and as such cannot be put to practical use. If this project were to be continued it would be by expanding the range of games that it can be used with, including networking for multiple clients to interact with each other and