Pneumonia is a form of acute respiratory infection that affects the lungs. The lungs are made up of small sacs called alveoli, which fill with air when a healthy person breathes. When an individual has pneumonia, the alveoli are filled with pus and fluid, which makes breathing painful and limits oxygen intake.
The aim was to take the dataset which contains 5,863 X-Ray images, divided into normal and pneumonia, and to construct a convolutional neural network that is able to classify whether the given x-ray image can be diagonsed with pneumonia or not.
The given netowrk utlizies transfer learning, the network architecture is based on VGG16. The network was able to reach an accuray of 94% on the data test set.
The following metrics are of the network's performance over the test set:
- Overall Accuray: 94%
- Overall Recall: 96.67%
- Overall Precision: 94.01%
- AUC Score: 93%
Below is 16 images pulled randomly from the data test set, each image is labeled with the its true finding as well as the model prediction:
