Inspiration
Hard to control presentations is cool. But technology like the Myo is expensive. Wanted to leverage image processing to do something cool.
What it does
Let's you control Google Slides using gestures directed at your webcam. Also includes an additionally functionality that overlays a live stream of user posts, similar to what Periscope does.
How I built it
By taking advantage of the APIs available to chrome extensions, I was able to control Google Slides with some background scripts. First, I accessed the webcam feed, and fed it through a web socket to a python based server. After decoding the feed, I leveraged the power of OpenCV to process the video frame, after which I sent back down data about the image sent to the server.
Challenges I ran into
There were a couple challenges. One was getting a feed to OpenCV, and then sending it back down to the client. Another was the issue of needing to have a web socket open (so that the processed feed could also be visualized in the browser), while also having to send back other information about each frame (in this case, which direction the gesture was in).
Accomplishments that I'm proud of
Even though what I wanted to build was somewhat ambitious, I was able to get something together that worked relatively well!
Log in or sign up for Devpost to join the conversation.