Activity is an IOS App that show how to train and deploy a Human Activity Model in your mobile thanks to CoreML and TuriCreate

Activity recognition is a really exciting field because it can associate a service based on what the user is doing.

Eventually, I'll associate activity with music.

I want to classify severals kind of activities :

Simple activities :

• walking
• running
• sitting
• jumping

Complex activities :

• metro
• car
• skating
• biking

I recorded my movements - accelerometers and gyroscopes - during each activity with my Iphone7 sensors.

My python project is here, I used Turi Create for the training part.

Machine Learning Architecture

The app is in Swift 5 with two targets :

• Activity : the app
• CoreActivity : the motion data handling and predictions with CoreML.

• CoreML
• DeviceMotion
• TuriCreate
• Building Activity Classification Models in Create ML
• Activity Recognition with Smartphone Sensors