Like many of us, I have spent more and more time in front of a computer in the last decade. I’ve always tried to be healthy and listen to my body, therefore I’ve been quite surprised when my trapezius muscles started feeling tight and sometimes even painful.
At first, I thought that maybe I didn’t stretch enough after exercising myself. But the discomfort occurred even when I didn’t do sport for several days. After mentioning it to a doctor, he immediately ask me if I often worked on a computer.
It was right on the mark !
Poor posture when sitting in front of a computer lead sooner or later to muscular tension if not pain. This is how I realised that the importance of a good sitting posture is rarely mentioned in our digital world.
Image Credit : https://blog.bufferapp.com/improve-posture-good-posture-science-happiness
So when I stumbled upon a GitHub project that aims at notifying slouchy people, it immediately kept my attention. Unfortunately, the software (in Python) wasn’t up to date and I didn’t manage to run it on my MacBook.
After spending a couple of hours trying to get it working, I said to myself :
Why not doing it as an application in the browser so that anybody could use it without having to struggle installing dependencies.
Hence is born the application SlouchyJS.
Like the original project I needed to use pictures of the user taken by the webcam. But the accuracy of the software weren't so great using old facial recognition techniques.
So I decided to use top notch deep learning to classify the user's posture as either standing straight or being slouchy. In upcoming articles I will detail each step I have taken to realise this web application from basic front-end javascript to convolutional neural network transfer learning.
So stay tuned if you want to learn more about the way AI is used in today's world ( and also how to solve your back pain ) !
Try it by yourself ! ( Still in development )
Original GitHub Repo ( seems inactive for some time )
Made with fun by Hugo Rosenkranz.
