Bedsore Patient's Pose Classification Based-on Human Skeleton

This is my first project that I'm working with Machine Learning .Hope you guys enjoy!

Preparation

for this project you should read the requirements software in requirements.txt first

in this project we have to use tf-pose-estimation or openpose for pose estimation software (if you don't this program will not be able to run)

Installation Guide

1. just git or download this project to your dir

git clone https://github.com/northnpk/OPEN-BS.git
cd OPEN-BS/

2. run

To run the code use

python3 01_OPENBS_Test.py

or

python 01_OPENBS_Test.py

I've try it on Python version 3 but on python 2 It could be run on.

run on video file

python3 01_OPEN-BS_Test.py \
    --model_path model/trained_classifier.pickle \
    --data_type video \
    --data_path data_test/video.avi \
    --output_folder output

run on a folder of images

python3 01_OPEN-BS_Test.py \
    --model_path model/trained_classifier.pickle \
    --data_type folder \
    --data_path data_test/path \
    --output_folder output

run on webcam

python3 01_OPEN-BS_Test .py \
    --model_path model/trained_classifier.pickle \
    --data_type webcam \
    --data_path 0 \
    --output_folder output

This program can work on CPU but if you need to use it on GPU

please read the requirement software from tf-pose-estimation

and I Recommently !! you should use Nvidia GPU (at least GTX 1060) and check about tensorflow build with CUDA and cuDNN versions support (Yes I have my own with RTX 2060 and have a lot of problem with tensorflow version and CUDA version)

Before you go please don't use this for commercial until I update this Project with higher accuracy.

And This project was supported by JSTP and NSTDA before use it please check LICENSE.txt

Thanks!