My idea is for a math tutoring platform that makes it as easy as possible for tutors to connect with students. A pain point for tutors are when students cancel at the last minute, or don't show up. A way around this is to conduct the tutoring sessions entirely through a video chat, but this complicates written communication. Rather than requiring the tutor and student to write math using a stylus, it would be awesome to use deep learning to allow both sides to "write" in the air with their finger or a pen.
Thoughts? Anyone know of something similar that already exists?
Transcribing motion via video would be amazing, think there could be bigger markets than tutoring. Have you already explored how this can be done?
There are definitely other applications of computer vision. I just picked one that I can relate to. I would be interested to explore other ideas once I get comfortable with the tools I'm working with.
I've been playing with open cv: https://docs.opencv.org/master/d1/dfb/intro.html. Haven't gotten to the point where I can track an object yet, but I'm pretty certain it is possible.