In this project, I create a model capable of classifying a chest x-ray as pneumonia or normal. Pneumonia is an infectious disease that inflames the air sacs of one or both lungs. It can be life-threatening for anyone who has it. In order to help physicians in their decision making, we are carrying out this project.
Several steps were followed throughout this project:
- Dataset creation (training, validation and test datasets).
- Data processing and data augmentation when necessary (the latter only for the training and validation datasets).
- Creation and training of convolutional neural networks using Tensorflow 2.0 (InceptionV3, ResNet50).
A comparison is performed at the end to choose the best model.