PneumoniaDetectionDeepLearning

Pneumonia detection using deep learning with Python and Tensorflow/Keras.

Dataset

Kermany, Daniel; Zhang, Kang; Goldbaum, Michael (2018), “Labeled Optical Coherence Tomography (OCT) and Chest X-Ray Images for Classification”, Mendeley Data, V2, doi: 10.17632/rscbjbr9sj.2

Link

Original task is a binary classification (normal/pneumonia):

data
└───train
│   └───normal
│       │   ...
│   └───pneumonia
│       │   ...
└───test
│   └───normal
│       │   ...
│   └───pneumonia
│       │   ...

I converted task to multi-class classification (normal / pneumonia_bacteria / pneumonia_virus):

data
└───train
│   └───normal
│       │   ...
│   └───pneumonia_bacteria
│       │   ...
│   └───pneumonia_virus
│       │   ...
└───test
│   └───normal
│       │   ...
│   └───pneumonia_bacteria
│       │   ...
│   └───pneumonia_virus
│       │   ...

Flask App

One can run flask app with:

FLASK_APP=app.py flask run