Inspiration
who doesn't want to type while doing the YMCA?
What it does
Types letters based on body position.
How we built it
Skeletal Tracking data from Asus Xition Pro Live piped into a tensorflow neural net.
Challenges we ran into
Setting up openNI for the Ausu Xition proved to be beyond trivial, and figuring out how to structure the data to feed into tensorflow.
Accomplishments that we're proud of
Training a neural net in such a short amount of time, understanding how to train a neural net using softmax regression, and getting to dance the YMCA all night long.
What we learned
How to work with skeletal tracking on Asus Xition, how to train a neural net, and how to structure position data into trainable features.
What's next for kinetic-keyboard
Support for the entire alphabet and emojis
Built With
- asus-xition
- c
- c++
- java
- machine-learning
- nite
- openni
- python
- sockets
- tensorflow

Log in or sign up for Devpost to join the conversation.