En | Ar

Projets de Recherche

Deep Learning Application on Chest X-Rays

Description :

The concept of machine learning, using algorithms to analyze and learn from data then take corresponding decisions, has been introduced to computer-aided diagnosis (CAD) of chest x-ray after being used for fifty years. As time progressed, deep learning stepped into making intelligent decisions on its own after creating a Neural Network. In a general vision, healthcare specialists started to feel the shortage of information given by 2D images, and longed forward 3D imaging of organs (ex: Computerized Tomography CT). However, deep learning can make use of 2D images, their labels and their corresponding CTs to learn and take decisions for future given 2D lateral images. In addition, deep learning should be capable of reconstructing 3D information from 2D images, and smartly diagnosing patients by analyzing 2D images. To be more specific, Thoracic diseases would be diagnosed more accurately when chest X-rays are supported and training X-rays with their corresponding CTs are available, in addition to a smart algorithm –to be developed- that would do the work.

Titulaire :
SAKR Georges

Contact USJ :

Projet présenté au CR, le : 01/11/2018

Projet achevé auprès du CR : 01/11/2021