Deep learning for digital pathology image analysis (practicals and lectures)

This course is now over-subscribed with a waiting list.

Overview

There is a confluence of ongoing revolutions in biomedical image acquisition, computational methods, and technologies for extracting and integrating relevant data in the construction of diagnostic and prognostic models of disease. In particular, deep learning methods are rapidly expanding the range and accuracy of tools for pathologists and researchers.

The aim of this course is two-fold: discuss applications of state-of-the art deep learning methods in digital pathology, and provide practical training in these methods. The course will consist of a half-day of lectures followed by an optional half-day of practicals. The practicals will be based on the PyTorch framework (please see course prerequisites below).

At the bottom of this page there is a link to proceed with registration for the practicals and lectures.

To register for just the lectures, please follow this link.

This event is co-hosted with the Swiss Digital Pathology Consortium (SDiPath).

Audience

This course is targeted to clinicians and researchers who are interested in discovering deep learning methods for the analysis of histopathological image data.

Learning objectives

Participants will be introduced to the relevant topics and state-of-the-art methods for deep learning in digital pathology image analysis. Those participating in the practical session will be able to:

• prepare a suitably formatted dataset for deep learning methods
• perform quality control of digital pathology slides
• understand the basics of working with the PyTorch framework
• perform segmentation and classification

Prerequisites

Knowledge / competencies

Participants should be comfortable working with Python and Linux/UNIX, having completed at least introductory courses in these topics (eg Python, UNIX). Participants should also have taken at least introductory training in general statistics.

Technical

Participants should bring a laptop with wireless connectivity. Windows users must have software installed for establishing remote ssh connections (eg MobaXterm).

Application

The registration fees for academics are 60 CHF. This includes course content material and coffee breaks. Fees for participants from non-academic institutions are 300 CHF.

Deadline for registration and free-of-charge cancellation is set is set to [08/03/2019]. Cancellation after this date will not be reimbursed. Please note that participation to SIB courses is subject to our general conditions.

You will be informed by email of your registration confirmation.

Venue and Time

University of Basel, Biozentrum, Klingelbergstrasse 70, Hörsaal 103 (morning lectures) and Seminarraum 104 (afternoon practicals)

• 9-9h10 - Welcome and introductions
• 9h10-10h10 - Raúl Catena, multi-layer tissue analysis
• 10h10-10h30 - Catered coffee break
• 10h30-11h15 - Tobias Sing and Pierre Moulin, "Pathology 2.0" and decision support in the clinic
• 11h15-12h00 - Andrew Janowcyzk, overview of deep learning and applications in research
• 12-13h15 - Lunch (on your own)
• 13h15-15h - Data formatting and quality control
• 15-15h20 - Catered coffee break
• 15h20-17h - PyTorch intro, segmentation and classification

Precise information will be provided to the participants in due time.

Additional information

Coordination: Geoffrey Fucile

You are welcome to register to the SIB courses mailing list to be informed of all future courses and workshops, as well as all important deadlines using the form here.

For more information, please contact training@sib.swiss.