Many people are left with poor fine motor skills due to an accident, or medical condition. This makes it difficult for the person to write and type and the only way to record text is to use some form of speech to text application. However, we wanted to provide people with such limited motor ability another option, one of being able to write with their limited skills.
What it does
We created a system that can record a person's hand movements while writing and convert it to text. We do this by adding an accelerometer to a pen (via an Arduino), and apply machine learning on the recorded data to determine what letter has been written.
How we built it
We used an Arduino with an accelerometer to record the hand motions while writing. We used the sci-kit learn python toolkit to apply machine learning on the data recorded.
Challenges we ran into
We were unable to implement the machine learning for every letter of the alphabet and account for periods and spaces due to time constraints.
Accomplishments that we are proud of
We were able to get a pretty stable hardware system in place and learned how to work with machine learning APIs.
What we learned
We learned how to work with real world data in machine learning applications.
Implement the system for all letters of the alphabet and account for periods and spaces while writing.