Malaria Model 🤖

Convolutional neural network that predicts if an image of a cell is parasitized by malaria or not.

Goals

Predict if an image of a cell is parasitized by malaria:

Parasitized

Uninfected

Results

Confusion matrix

[[1165  135]
 [  45 1255]]

Classification report

              precision    recall  f1-score   support

           0       0.96      0.90      0.93      1300
           1       0.90      0.97      0.93      1300

    accuracy                           0.93      2600
   macro avg       0.93      0.93      0.93      2600
weighted avg       0.93      0.93      0.93      2600

Exploratory data analysis

Loss

Accuracy

Running

To run the model first create a new Python environment and activate it. I'm using Anaconda for setting the python version that pipenv should use to set up the environment. The command bellow will automatically setup the environment with conda and pipenv:

make env

Now install all the project dependencies:

make install-all

And run the model:

make model

Warning

Dont run make model without deleting storage/malaria-model.keras, this will cause train/test data over fitting.

After running you model, it will be saved inside storage/malaria-model.keras. To just run your recent created model and predict a random value from our data set, use the following script:

make predict 

Warning

In case you have deleted the storage/malaria-model.keras, remember that to get make predict working you need to run make model first to create it.

To run TensorBoard with the latest created version of the model within this repository run:

make board