neckTech

Front Screen
Home Screen
Calibration Screen
GIF
Demo Of The Machine Learning Model

Inspiration

As COVID-19 ravaged the world, people were told to isolate themselves, and stay inside. Many people started working from home. People started working in their beds, couches, and kitchen tables, often sitting for hours at a time without any breaks, causing many neck and back issues due to bad posture when sitting in those positions. We want to help correct that.

What it does

Ensures that your posture is correct and encourages you to do some stretches given a timer,

How we built it

Our application is really lightweight and built using electron.js, which means that our application can also run on your browser. We make use of ML5 on Tensorflow.js for our machine learning base, and we used Photon as a basis for our UI.

Challenges we ran into

We initially had difficulty in connecting the two neural networks to our electron application. We handled it step by step and worked through each view. This way we were able to make it work eventually.

Accomplishments that we are proud of

We are proud of building this application in 24 hours and making it fully functional within this timeframe. We are also proud of having a great developer coordination / collaboration.

What we learned

This was the first time all of us worked on an electron project, but more importantly, with machine learning. We learned a lot about classification, which allows us to take inputs, specifically our body points, and use that to predict whether or not you have good posture. We've learned a lot about Machine Learning, and how to set up Neural Networks and this hackathon was fun and a great learning experience.

What's next for neckTech

This project is fully extensible and will include more exercises and help support seating ergonomics in different ways. After this hackathon, we’d like to work on fine tuning our models and make it available for all students and employees around the world who are now working online.

Built With

electron
javascript
jquery
ml5
p5
photon
posenet
tensorflow

Submitted to

nwHacks 2021
- Winner Honorable Mention

Created by

I jumped around everywhere, but my focus was on designing and integrating our ML infrastructure. I came up with the ML design infrastructure for the two NNs and I helped connect the UI with the ML.

I really enjoyed doing this Project with my team and I'm very proud of what we've managed to build in 24 hours!

Devam Sisodraker
I helped out with training the machine learning model.

Jack He
I engineered and implemented the machine learning model using a Neural Network and integration with frontend

Liang Liu
I worked on integrating the electron.js application with other functionalities. I gained experience in using the neural network for exercises. I also worked on the design aspects of the UI/UX. This hackathon was a great learning experience and it was fun hacking with the team!

Vishal Desh
Computer Science - Roboticist - University of British Columbia