Bookmark and Share

Advancing AI-Driven Oculomics for Eye and Systemic Diseases


University of Oulu


Learning objectives and contents: This short course explores the intersection of artificial intelligence (AI) and ophthalmic imaging, focusing on addressing eye and systemic diseases through advanced AI methods. It equips students with foundational knowledge, practical skills, and ethical awareness to contribute to AI-driven healthcare. The course begins by introducing ophthalmic imaging techniques, such as detecting diabetic retinopathy and glaucoma, and their links to systemic conditions like diabetes. Students will then study advanced segmentation methods, from convolutional neural networks (CNNs) to foundation models, emphasising their evolution and practical use in medical imaging. A module on diagnosis, prognosis, and risk prediction highlights AI's role in detecting diseases and predicting progression, integrating imaging and clinical data to develop robust models. Ethical and practical challenges, including bias, reliability, and privacy concerns, are also addressed. Hands-on learning is a key component, with a retinal image segmentation case study and a group project on fundus image classification. These practical exercises allow students to apply theoretical knowledge to real-world datasets, tackling the complexities of medical AI projects. By the end, students will be well-versed in cutting-edge AI techniques and their applications, ready to make impactful contributions to healthcare innovation.

Pre-requisite: Basic concept on algorithm, image processing, deep learning, linear algebra and calculus.

Tentative topics:

  • Introduction to Ophthalmic Imaging and Common Eye and Systemic Diseases (one lecture)
  • Segmentation Methods from CNNs to Foundation Models (two lectures)
  • Diagnosis and Prognosis/Risk Predictions (two lectures)
  • Safety, Reliability and Ethics of AI (one lecture)
  • Case Study on Segmentation of Retinal Images (one lecture)
  • Group project on Fundus image Classification (one lecture)

Key methods:

  • Segmentation Models
  • Diagnosis, Prognosis and Risk Prediction
  • Domain Adaption
  • Self / Weakly/Semi-supervised Learning
  • Foundation Models

Amount of contact teaching hours:

  • 8 hours (including one hour for introducing group project and assessment)

Bio of the lecturer:

Prof Yalin Zheng is Professor of AI in Healthcare at the University of Liverpool, UK. He is a computer scientist working at the interface of artificial intelligence and healthcare, specialised in the research and development of artificial intelligence, biomedical image processing and analysis techniques. He has been involved in funded research of over £20 million from the UKRI, NIHR, Wellcome Trust, EU Horizon and industry. He has leadership roles as project PI delivering successful projects by working effectively with different stakeholders (academics, clinicians, public, funder representatives and industry). His collaboration with clinicians has made significant impacts on patient care. His recent AI work has led to a new University spin-out, AI-Sight Ltd, for screening of diabetic retinopathy. He has published over 300 papers in high-quality medical and engineering journals and conferences. He is an editorial board member of several journals and chaired/co-chaired several national/internal conferences.


Back

Course dates
26 February 2025 - 28 February 2025
Course organizer
University of Oulu
Place/Venue
Pentti Kaiteran katu 1
90570
City
Oulu
Country
Finland
Workload
2
Link
https://www.oulu.fi/en/education-search/advan...