neurostratum
- 9 Devlogs
- 6 Total hours
an accurate gait analysis app to fix postures and the way of walking
Ship #1
Pending review
@yaseenmerollin
on
neurostratum
AFTER DAYS OF RESEARCHING ABOUT MEDIAPIPE AND THE MATHS BEHIND TRACKING WALKING DATA AND 6 HOURS OF CODING I CAN PROUDLY SAY THAT I FINALLY FINISHED MAKING NEUROSTRATUM!!
It tracks human movement through a camera and turns raw gait/walking data into clinically-styled metric based reports. It has three core modules:
1.live gait tracking and logging:
This module lets you track your gait data and you have an option to save this data into a csv which then you can upload to the third module to get a pdf report with graphical demonstration aswell. the response in the pdf depends on your gait data.
2.Live posture tracking:
this module calculates your posture alignment and according to that it tells you if your posture is balanced or imbalanced. Sounds simple but is a huge factor for good posture!
3.pdf report of your gait data
This module takes your csv data of your walking, ensures if the metrics are normal or abnormal and then depending on your input gives an appropriate clinically-styled metric based reports. The pdf also gives you advice depending on wether your walking is good or bad.
Known limitations:
-Posture tracking has a noticeable delay (around 3 seconds) between movement and status update which is a target for future optimizations. -Gait metrics which are recorded live still need a multi step detection algorithm to make the system have depth awareness of the person its tracking to have an accurate metric report.
How to run it this project and test it locally:
Prerequisites
Node.js (v18+ recommended)-> includes npm: download here
Python 3 (latest) : download here
pip packages: numpy, pandas, matplotlib, fpdf
VS Code (recommended, not required)
download all the files as a zip (or clone it),unpack it and then install all the requirements for the libraries, check if you have npm installed and then type “npm start” in the project terminal, click the localhost link and youre good to go!
@yaseenmerollin
on
neurostratum
40m 27s logged
dev log 9- trying to deploy this app has been a nightmare. it is almost impossible to do so without paying for the web services so instewad ive decided to add a demo video showing how the project works with the files they can download and run by themselves
@yaseenmerollin
on
neurostratum
1h 2m 34s logged
dev log8- I FINALLY FINISHED MY GAIT DETECTION APP!!! I’m insanely happy with how this turned out and now all thats left is to fire up my local nodejs server test to see its accuracy, beautify things and then FINALLY i’ll learn how to ship my project (since its got a python backend which i’d love to get help from yall on how to ship) and we’ll move on to the next project!!
@yaseenmerollin
on
neurostratum
50m 13s logged
devlog 7 - just finished the program which checks your posture! it uses our gait engine aswell because it checks a few other metrics including the shoulder alignment, finally all i need to do is create the 3rd and final program which has an option to take in the walking style measured in a loong time (saved as csv) and report given to us as a pdf. Then i make the index.html and try my best to ship the project!
@yaseenmerollin
on
neurostratum
38m 50s logged
Devlog 6- finalised my pdf generation for the gait analyzer this morning and safe to say this was a PAAIN to make i had to learn so many new terms it was crazy anyways the pdf generator basically takes your gait data (cadence, stride length,step length, double support time) in a form of csv (which my gait analysing engine tracks real time and creates) and gives you an automated response of all the data on wether your walking style is good or bad, ive got only a few more html files to make and finally i can ship this project!
@yaseenmerollin
on
neurostratum
43m 59s logged
43 mins later and alot of changes had to be made to all of my files since im updating them as im moving forward with each new file. now i’ve been working on the csv data to pdf file converter in python and hopefully it goes smooth!
@yaseenmerollin
on
neurostratum
28m 31s logged
After hours of understanding how html works i finally made a html page for tracking the gait of a person whose walking! took me long enough lol now im gonna move on to other parts of the program!
@yaseenmerollin
on
neurostratum
31m 3s logged
Devlog-3 : The python engine: so after a long 26 minutes i finallyy finished the python engine which basically tracks the legs of the person its viewing and calculates the distance the ankle, heel etc are from each other and then with some arc tan finding angle yabadabadobee, it finalllyyy worked. now i’ve started on the html part and safe to say this is a pain in the butt. but away we go!
@yaseenmerollin
on
neurostratum
29m 38s logged
because i had to change the folder name, the hackatime time checker messed up and ruined the timing i had actually done, which was for 57 minutes. but nonetheless the initial writing for server.js is complete! now ive moved on to making a gait engine which basically tracks the way you walk. this will be the core part of the project and as im making it im able to learn a little about maths aswell lol
@yaseenmerollin
on
neurostratum
23m 52s logged
DEVLOG 1: i learnt how to set up a server with node.js and it seems quite complex but turns out that its just this obstacle i need to cross to start locking in on the actual app!!! And just by learning how to initialise your server with server.js, i’ve understood how the backend will be calling the specific apps! this is going to be a looooong project andI’m excited to share the journey with yall!!
