Love it or hate it AI is one of the most important technologies now in days, and as you might expect one day I had to take a class on it and make a project with this technology.
AI for wearables class sound like a good class to use the concept that I lend on my own about AI and I love the apple watch, I was thinking to make an app using this device, the perfect formula to kill two birds with one stone.
The Proyecto was straightforward, use native development to create an app for IOS that uses a motion sensor on the phone and implements a TensorFlow model that is already trained to identify 20 “common” gestures. These gestures will trigger a message. This app is for people with restricted mobility who can’t talk but can make repetitive moves.
The professor needs a native app because of some real-time and permission issues that happened in previous semesters.
For this project there were two leaders, Rafale and me, he was in charge of the IA part because he had more experience he had to dealing with the model, and making some modifications for the data type and sizes, on the other hand, I was working on the user interface, managing the libraries, making TensorFlow work on our app.
The idea for the UI was to keep it simple, clean, and mainly have big and few buttons, because the main user will be a disabled person with restricted mobility.
For the first part tap, I thought, if I was a teenager, I probably will like to have something to show and explain how the app works to my friend, so I add 3 speed miters that show the acceleration (gravity) on the 3 axes, this way the person can test if the sensors are working, play and use it to explain how it works.
The second one is the main one, it only has a pause/play button and a timer that count the recording (of the action) time.
And finally, on the third tap the user can personalize the phrases associated with each action, this phase automatically saves after modification.
Something that I loved about this project was that I work with people that had a different specialty (they were computer science MS and I was a Biomedical MS) than me and although we speak the same language averting was different and we communicate whit input and outputs of functions, many time some members of the team hand no idea of how a function work but they knew what to give and what to expect out of it.