A Machine Learning tool to diagnose thorax disease from X-ray images
Project Abstract
The 2010s saw the advent of ML techniques in medical diagnosis and this project aims to add to this growing body of research. There is often a delay between an X-ray being taken and diagnosis, which can possibly lead to more serious or fatal outcomes. My project aims to address this: further research in this area could save lives.The core of our strategy is a Neural Network algorithm that aims to diagnose up to eight possible diseases trained on a dataset of over 100,000 X-ray images of over 30,000 unique patients. Our project could identify issues in a few moments before the data would be analysed by a physician.The approach was successful: accurate enough to be added into routine checks, but never enough that it could ethically replace a physician?��s work, rather, multiply what they could do.In conclusion, this project serves as one of many attempts at testing the boundaries of what Machine Learning can do in medicine.
Keywords: Machine Learning, Deep Neural Networks, Medical Diagnosis
Conference Details
Session: Poster Session A at Poster Stand 69
Location: Sir Stanley Clarke Auditorium at Tuesday 7th 13:30 – 17:00
Markers: Benjamin Mora, Hans Ren
Course: MSci Computer Science, 3rd Year
Future Plans: I’m continuing studies