- This event has passed.
SICSA Seminar – Generative AI for Medical Computer Vision
Join us on Thursday, 30 April for an AI & Data Science Seminar.
Dr Hazrat Ali of University of Stirling presents
Generative AI for Medical Computer Vision
Abstract
Before synthetic data became popular, data scarcity was considered a major bottleneck in training AI (deep learning) models for medical imaging. Generative AI enabled the AI research community to feed the data-hungry deep learning models by generating synthetic image data. This talk will explore the potential of Generative AI, particularly Generative Adversarial Networks, in medical imaging applications. The talk will cover some of the advances in; (1) atherosclerotic disease transformation, a project that I undertook at the Umea Biomedical Engineering R&D of Umea University Hospital, (2) Super-resolution of medical images, and (3) Recent shift toward Neural Diffusion Models.
Dr Hazrat Ali is Lecturer in AI and Data Science at University of Stirling where his research portfolio spans Generative AI, Medical AI, healthcare, and computer vision, with a specific focus on generative adversarial networks for medical imaging, deep learning for ultrasound medical imaging, and AI for healthcare. Dr Ali developed a new method for disease transformation in ultrasound images, as part of the VIPVIZA project at Umea University, Sweden. In addition, he has developed open-source datasets, including a dataset of natural scene images with Urdu text, the UHaT dataset (the first dataset on Urdu handwritten characters), and a speech corpus for Urdu. These datasets, available on the Kaggle platform, have proven invaluable for AI-oriented tasks and have contributed to advancing the field.
He has worked on multiple research positions, including working as a senior researcher at Umea University Biomedical Engineering R&D Center in Sweden, and Hamad Bin Khalifa University in Qatar.