Md. Hasnat Riaz

Research Title: Machine Learning and Deep Learning Approaches for Ocular Disease Prediction

Supervisor: Dr Maitreyee Roy

Co-supervisor: Dr Andrew White and Dr Sundaresan Raman

Research

Abstract

Machine Learning (ML) and its subset, deep learning (DL) techniques, are currently showing great promise in the field of healthcare. They are leading a revolution in the diagnosis based on image interpretation, disease prediction, and risk assessment. In ophthalmology, Artificial Intelligence (AI) is becoming increasingly prevalent, aiding in inspection, image analysis, rapid diagnosis, and treatment recommendations for eye conditions.

In the future, there is an anticipated dramatic increase in ocular diseases due to the aging population. In such a scenario, the early identification and proper management of ocular diseases represent a cost-effective and fruitful approach to prevent blindness caused by conditions such as diabetic retinopathy, glaucoma, cataract, retinopathy of prematurity (ROP), age-related macular degeneration (AMD), and numerous other diseases. This approach serves to preserve vision and enhance the quality of life.

According to the World Health Organization (WHO), at least 2.2 billion people worldwide have near or distance vision impairments. In most cases, vision impairment could have been prevented or can still be addressed. There is an urgent need for the immediate and automatic detection of diseases to reduce the workload of ophthalmologists and prevent vision damage in patients. ML and DL in ophthalmology are invaluable, as they accelerate the diagnostic process and reduce the human resources required.

Recent research in bioinformatics has witnessed the development of ML and DL algorithms for medical image interpretation, including ocular images. DL has proven to be highly efficient in identifying crucial clinical features for the diagnosis of ocular diseases. ML and DL-based models have also been explored for fundus photographs (FP) and optical coherence tomography (OCT) images. FP provides enlarged images of retinal tissues, which are well-suited for monitoring, diagnosing, and treating ocular diseases, such as diabetic retinopathy (DR), glaucoma, ROP, and AMD. Meanwhile, OCT provides critical morphological information, playing a significant role in the assessment of DR, AMD, and glaucoma.

However, it's worth noting that the interpretation of an FP or OCT image can vary among ophthalmologists, which can be problematic for patients who need to make clinically sound decisions.

This PhD research aims to identify the most effective machine learning methods for pattern recognition, decision-making, and the diagnosis of ocular diseases with minimal human intervention through ocular image analysis.

Biography

Md Hasnat Riaz earned his MSc and BSc degrees from the Department of Computer Science and Telecommunication Engineering at Noakhali Science and Technology University, Bangladesh. Following his academic journey, he spent seven years as a faculty member within the same department at the same university. Furthermore, he gained professional experience as a software engineer at Leads Corporation Limited in Bangladesh and successfully completed an internship at Infosys in Mysore, India. 

Education

PhD candidate at UNSW Sydney (Current)

MSc. in Computer Science and Telecommunication Engineering from Noakhali Science and Technology University, Bangladesh, awarded in 2016.

BSc. in Computer Science and Telecommunication Engineering from Noakhali Science and Technology University, Bangladesh, awarded in 2012.

Awards

Prime Minister Gold Medal Award, Bangladesh.

Vice-Chancellor’s Gold Medal Award, Bangladesh.