WorkIN

WorkIN

Inspiration

What it does

WorkIN is designed to allow users to easily and efficiently do their fitness regimes inside their homes. Users would be able to carry out their routines without any need of external training/help.

How we built it

We use pose estimation models using tensorflow.js and feed them into ml5.js allowing characterization the yoga postures and then cross examine the poses with the current indivduals posture and allow the timer to start if and only if the correct posture is achieved. Transition from one yoga pose to another following the yoga routine.

Challenges we ran into

Tedious training procedures for training the model Proper implementation of time running when pose is achieved

Accomplishments that we're proud of

Running of the web app proper pose detection models flow of the website

What we learned

Posenet p5.js tf.js modules neural nets on js environment

What's next for WorkIN

We aim to keep several analytical tools to help the user figure out, in what areas, he/she needs, to work upon. We can also have a video description of the correct poses being done, and also include a mobile app.