Rahmat Izwan Heroza Ibrahim

Deep Learning Methods for Medical Image Analysis

This project aims to develop new methods for medical data analysis using deep learning. The problems to be investigated include how to train deep learning models effectively with imbalanced datasets and how to achieve ~100% sensitivity with deep neural networks for medical diagnosis. This project will likely make original contributions via proposing new methods for data augmentation, new deep learning model architectures and information fusion methods for improving the accuracy and sensitivity

Supervisor: Professor John Q. Gan , Dr Haider Raza