Overview
As an exploratory dip into the world of AI, I had set out to create a binary classifier that was trained on a kaggle dataset that contained cells that were infected with malaria, and cells that were considered to be uninfected.
Design
The design for this was fairly simple. I had pulled a dataset of cell images containing both infected and uninfected malaria cells and assigned a label "0" or "1" to each of them. These were then
each cell to represent their state to then turn them into a numpy array.
the training step had a simple architecture, with 3 convolutional layers that had a pooling layer between each step, and then flatten the output before creating a fully connected layer for my two classes.
Results
At the end of this, I was able to achieve an overall accuracy of about 95% on recognizing infected cells.