The personalized AI posture practitioner that prevents cervical & lumbar vertebral diseases from sedentary lifestyle using deep learning and computer vision
-
Presentation: Investor Version & Developer Version
-
Presentation Slides: Investor Version & Developer Version
Python version > 3.3
conda create situp_env
conda activate situp_env
- install all requiment according to AI module
- copy paste all files in sitUpBackEnd to folder of pose-estimation-detection cloned from AI module
- under conda environment, run
$ cd pose-estimation-detection
$ python main.py
- go to ionic folder and start the app
$ cd sitUpFrontEnd
$ npm i
$ ionic serve
enjoy the app!
Use tf-pose-estimation
cd sitUpBackEnd
For download of dependencies, refer to https://medium.com/@gsethi2409/pose-estimation-with-tensorflow-2-0-a51162c095ba;
For download of dependencies, refer to https://github.com/satyaborg/pose-estimation-detection;
mkdir result
Note directory result is under sitUpBackEnd;
Move sit-up.py and sit_analysis.py to the folder "tf-pose-estimation"(For mac users)/"pose-estimation-detection"(for windows users) which should be downloaded after step 1;
Simple test! go to the subfolder just created and run with webcamera: (for mac users): python sit-up.py --os=macos --direction=front
or
(for windows users): python sit-up.py --os=windows --direction=front
;
For the argument direction, you can choose front, side45 or side90;
Note when specifying side45, for the purpose of accuracy, please make sure camera is at 45 angle of your front;
Try the demo vison with only ionic
Here are two video demos
$ cd setUpFrontEnd
$ npm i
$ ionic serve
Then go to profile and hit the demo
button