This week I did research on the different technologies we should use to implement our project. Specifically, I looked into parts for the Arduino including the Bluetooth module, accelerometer, and projector. I also looked into how these parts would fit together and if there was any discrepancy in the parts we ordered communicating with each other.

I thought more about specifics regarding our implementation, especially on the Body Expression classifier. I originally thought using Open Pose would be a good idea for our project design, but because OpenPose needs to use an Nvidia GPU / a non Mac, we would need to relay our photo/video data to a separate computer to process.

This seemed inefficient to me, I wanted this project to be more of a standalone attachment that just required the user’s Mac laptop. I presented this concern to my team and we decided there might be better solutions. After doing more research, I thought OpenCV would be a better option for our use case. Instead of detecting body movements, we could train a classifier to detect hand motions. This is more lightweight than OpenPose, so it would make the attachment a standalone attachment the user could use.  Also, since it is a trained classifier rather than judging based on previous movements, this method would potentially be more accurate than the OpenPose idea.

With all this research regarding implementation, I helped work on the overall system design and plan for our budget/which parts we want to buy:

I believe we are on track. For next week, we need to finalize our system design and order the parts to create our device.