incorrect sitting position is one of the most important reason causing the back pain, especially we are the student and have to stay in front of desk for whole day to study. We are suffering at back pain and that makes us crazy.

We use our the acceleration transducer in the mobile phone to collect data. And then build the mathematical model to classify different postures. And the client side will receive the real time classification result, if the result is one of the incorrect results, the client side will receive a notification.

How we built it: we use acceleration transducer to collect data; build a data GUI tool to monitor different incorrect postures patterns; apply LSTM to building the model and deploy the app to our cellphones by react-native.

Challenges we ran into: collect data, build pipeline

Accomplishments that we're proud of: intuitive usage, beautiful UI, AI

What we learned fast prototype

What's next for Sitting companion robust model

Share this project: