Pneumonia Detection Using Convolutional Neural Networks
Challenge
The aim of this work is to develop an efficient neural network to perform binary classification on X-ray images.
This work implement the architecture proposed in the paper "An Efficient Deep Learning Approach to Pneumonia Classification in Healthcare, Journal of Healthcare Engineering, 2019, Okeke Stephen et al".
Dataset
The dataset is organized into 3 folders (train, test, val) and contains subfolders for each image category (Pneumonia/Normal). There are 5,863 X-Ray images (JPEG) and 2 categories (Pneumonia/Normal).
You can found it on kaggle at https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia
Convolutional Neural Network
In Okeke Stephen et al., authors proposed the following architecture:
After reimplementing this model with the Keras Sequential API, we obtain the following architecture:
Results
After model testing, we obtain the following results:
Which leads to an accuracy of 0.89.