Using Chest-X Ray images from a very well known Kaggle Dataset (https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia), we can use Deep Learning models to detect Pneumonia.
Healthcare is an important part of the current technological revolution, with more and more fields like Deep Learning applying methods to this versatile and pivotal field. Pneumonia is a very common infection with more than 10 million cases annually. Detecting this infection at an early stage using a simple chest x-ray scan could prove to be a game-changer.
This project uses Tensorflow 2.1.0 and Keras 2.3.0. The same code might not work with Tensorflow 1.x.x.
This project was executed in the Google Colab environment, with a GPU configuration. Each epoch during training took around 45 seconds, which may differ everytime. The Kaggle API was used to import the dataset into the Colab environment.
This project used Transfer Learning techniques, with the base model being ResNet50 (https://keras.io/applications/#resnet), and the model has 50 layers.
After a mere 10 epochs of training, the model gets an accuracy of >80%. Some small tweaks to the architecture should bring it above 90%.
Contributions are open from everyone.