The aim of this project is to apply deep learning to the entire X-ray processing chain, from acquisition to image retrieval, from segmentation to disease prediction.
Chest pneumonia caused by bacterial or lung infection can cause life threatening consequences and, in most cases, leads to death. As a result, it is vital that diagnosis is carried out at an early stage to minimize any risks. Recent advancements in artificial intelligence and medical imagining have paved the way automation systems to be developed, capable of diagnosing x-rays, thus simplifying the pneumonia detection process for radiologist and other medical experts. The aim of this study is to develop and compare various models to help identify the chest x-rays (CXR), classifying them as either Normal (healthy) or Pneumonia (unhealthy). To achieve this, four existing state of the art Machine Learning (ML) models have been used. Experimentally results showed that Deep Learning (DL) techniques can be used to successfully classify CXR images, using DL based on Convolutional Neural Networks (CNN) (Mamlook, R., 2020) with the greatest accuracy achieved being 85%.
- Increase productivity by reducing labour intensive and repetitive tasks of analysing medical images.
- Provide faster diagnosis.
- Provide more accurate analysis than radiologists.