GitHub - arunkv1/DiabeticRetinopathy: Doctors always need imaging and computer references to diagnose severity of Diabetic Retinopathy. This is a model trained with over 16gb of images of the retina from kaggle and performs a multi-class classification to determine how severe the condition is. This model works at about 97% accuracy.

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 3 Commits
crossVals.txt		crossVals.txt
diabeticRetinopathy.py		diabeticRetinopathy.py
paramTuning.py		paramTuning.py
readme.txt		readme.txt
videLink.txt		videLink.txt

Repository files navigation

Running the Project:

To run the project, all you have to do is change the path to the file name in all the python files. The current file name is on our aws s3 bucket, so just change the file name to the one on your own personal computer. Also, in order to run the data bricks package's functions, you will need to download the necessary versions of the packages. The DeepImageFeaturizer will only work with tensor flow 1.6, and keras 2.5. In addition to this, you may have to download other required pip libraries and the current versions to make the code work. We created a conda environment on our EMR cluster, and downloaded all the required packages and correct versions for it to run perfectly. Once the correct versions are installed from the list in this readme file. Then the code will run perfectly. If you are having trouble running the file, feel free to reach out to us. Our emails are at the end of this readme file. 

---- The Required Pip Packages To Run The Code ----

Package            Version  
------------------ ---------
absl-py            0.11.0   
appnope            0.1.2    
astor              0.8.1    
backcall           0.2.0    
bleach             1.5.0    
certifi            2020.6.20
cloudpickle        1.6.0    
coverage           4.4.1    
decorator          4.4.2    
gast               0.4.0    
grpcio             1.34.0   
h5py               2.7.0    
html5lib           0.9999999
importlib-metadata 2.1.1    
ipykernel          5.3.4    
ipython            7.9.0    
ipython-genutils   0.2.0    
jedi               0.17.2   
jupyter-client     6.1.7    
jupyter-core       4.6.3    
Keras              2.1.5    
Markdown           3.2.2    
nose               1.3.7    
numpy              1.16.4   
olefile            0.46     
pandas             0.19.1   
parameterized      0.6.1    
parso              0.7.1    
pexpect            4.8.0    
pickleshare        0.7.5    
Pillow             4.1.1    
pip                10.0.1   
prompt-toolkit     2.0.10   
protobuf           3.14.0   
ptyprocess         0.6.0    
py4j               0.10.6   
Pygments           2.2.0    
pyspark            2.3.0    
python-dateutil    2.8.1    
pytz               2020.4   
PyYAML             5.3.1    
pyzmq              20.0.0   
scipy              1.4.1    
setuptools         40.2.0   
six                1.10.0   
sparkdl            0.2.2    
spyder-kernels     1.9.1    
tensorboard        1.6.0    
tensorflow         1.6.0    
tensorframes       0.2.9    
termcolor          1.1.0    
tornado            6.1      
traitlets          4.3.3    
wcwidth            0.2.5    
Werkzeug           1.0.1    
wheel              0.36.0   
wrapt              1.12.1   
wurlitzer          2.0.1    
zipp               1.2.0

About

Doctors always need imaging and computer references to diagnose severity of Diabetic Retinopathy. This is a model trained with over 16gb of images of the retina from kaggle and performs a multi-class classification to determine how severe the condition is. This model works at about 97% accuracy.

Readme