Thesis First Play Test

The first play test took place in the BallRoom.

I tested ml5 vs. OF HaarFinder vision model in a dim light situation, on a mac and a windows laptop.

ml5 on the web was great at catching face mesh regardless of the light situation. But there was a limitation of 1 face that it can detect at once.

On the other hand, OF model did not do too great with faces that are far away. However, it captured multiple faces & were great at detecting faces & dropping the detection when the user is looking away.

Tasks next up is to first, get the algorithm down to drive a motor & leds on a laptop, then to put that on a pi.

Got feedbacks that:

-I might need an attention grabber

-I would need to playtest this method with the full output soon.

-If I were to change Thesis subject, it should happen now.