报告题目：Revolutionizing Pathology Diagnoses: A Journey from Immunology Microscopy to Fully Automated Imaging

时间：2023年12月5日下午2:30

地点：大学城校区B8-报告厅

报告人：Brian Lovell 教授

Abstract:

Pathology plays a crucial role in diagnosing diseases worldwide, accounting for 70% of all diagnoses. However, the field faces significant challenges, such as a shortage of senior pathologists and an aging population. To address these issues, a transformative project was initiated to modernize immunology microscopy through the integration of digital imaging and computer vision.

Traditional microscopy techniques in pathology had seen little change over the past 50 years and remained largely not unlike the pioneering work of Louis Pasteur in the 19th century. The initial clinical project focused on digitally processing the ANA-IFF immunology test, which was successfully implemented in 2011, albeit covering only a fraction of the workload. This achievement led to notable benefits, including substantial cost savings in consumables, amounting to $150,000 in reagents and $25,000 in glass annually, and significantly accelerated diagnosis.

Building on this success, the research team secured two ARC grants and a Fellowship, enabling a dedicated effort towards designing a fully automated system capable of handling all microscopy processes and analyzing results. The primary challenge lay in achieving fast and 100% reliable automated image capture, particularly when dealing with challenging specimens at very high magnification, using a x100 oil immersion lens.

This talk will delve into the journey of how innovation and automation in pathology have the potential to revolutionize microscopy, making it possible to image almost anything fully automatically. The presentation will highlight the milestones reached and the transformative outcomes achieved in bridging the gap between traditional microscopy and cutting-edge digital imaging using AI.

Introduction:

Brian Lovell ，昆士兰大学电气工程与计算机科学工程、建筑与信息技术学院教授，昆士兰大学高级监控小组项目负责人，电气和电子工程师学会高级会员，国际电子工程师学会研究员，亚太人工智能协会研究员，自1998 年以来一直是国际模式识别协会理事会中具有投票权的澳大利亚成员。主要研究深度学习的人脸识别、利用深度学习进行物体检测、合成人脸和图像生成。