AI in healthcare will always be the reason I started to dig in this field of Artificial Intelligence. The scope is huge and the impact you make in the world, too. Making a prototype to solve medical problems will always be a reason to be proud of.
It’s unbelievable how AI is improving the healthcare field, specifically in medical diagnosis. AI will improve the way Doctors diagnose and treat diseases. It’s not a competition but an opportunity to join forces!
This time, detecting Pneumonia in Chest X-Ray images, is a great experience. I will detect Pneumonia in Chest X-Rays: Using a Convolutional Neural Network with Tensorflow.
According to the American Lung Association, Pneumonia is an infection that inflames your lungs’ air sacs (alveoli). The air sacs may fill up with fluid or pus, causing symptoms such as a cough, fever, chills and trouble breathing.
Most common Pneumonia symptoms are:
Cough, which may produce greenish, yellow or even bloody mucus Fever, sweating and shaking chills Shortness of breath Rapid, shallow breathing Sharp or stabbing chest pain that gets worse when you breathe deeply or cough Loss of appetite, low energy, and fatigue Nausea and vomiting, especially in small children.
I will use the Chest X-Ray Images (Pneumonia) Dataset. You can find this dataset at Kaggle. It’s organized into 3 folders (train, test and val sets) and contains subfolders for each image category (Pneumonia/Normal). There are 5,863 X-Ray images (JPEG) and 2 categories (Pneumonia/Normal).
Chest X-ray images were selected from retrospective cohorts of pediatric patients of one to five years old from Guangzhou Women and Children’s Medical Center, Guangzhou. All chest X-ray imaging was performed as part of patients’ routine clinical care.
My model achieved an accuracy of 80% on test data with 78% recall. Now you can classify the images into 2 categories: Normal and Pneumonia by using this model.
You can get the dataset here:
https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia/
The project is broken down into multiple steps:
Analyse and preprocess the image dataset .Creating an image classifier using dataset .Use the trained classifier to predict image content. Everything you need to recreate this project is on the jupyter notebook.