With the rapidly increased number of individuals around the world being forced to “work from home”, we are aware that many of you have far from ideal setups for your working day, with a lack of office-made chairs and desks as you spend the majority of your day at your work device. Naturally this leads to significant discomfort as you might find yourself gradually slouching further and further into your desk over the course of the day. Not only does this lead to immediate pain in the form of a strained neck or a sore back, but further down the line this can lead to even more significant health impacts. The main difficulty with this is that you may not even notice just how poor your posture is when you are seated for extensive periods of time!
Therefore we wanted to create a system which notifies you when it detects a poor posture. We are also incorporating a system which provides you with real-time visual feedback to guide you back to a “good” posture.
We are excited to announce the first full release of PosturePerfection! We have tweaked a few things since v0.4.0 to make your user experience even better, and we have added some guidance on how to use our system.
Set an ideal posture
Receive constant feedback on your actual posture
Receive notifications on multiple Linux devices when your posture becomes bad
PosturePerfection is entering the final phase before the first full release! We've updated the user interface to improve usability and included useful hints on the system's current state. Also added in this release is the notification system, whereby users can run a client on a separate Linux machine to receive notifications from the machine running PosturePerfection
Note: To use the notification receiver client you must build from source. This will be changed in the next release.
Feast your eyes on the newest PosturePerfection release, revising the real-time output. The live video input stream is now passed through the pipeline performing pose estimation and posture estimation before displaying the video feed with the detected posture displayed. A settings page has also been included which allows some configuration of the pipeline.
You can find this release under on our GitHub here.