The Image Analysis Collaboratory and Nikon Imaging Center are running a workshop: Segment Anything (SAM) in Python. SAM is a deep learning-based segmentation model introduced by researchers at Meta. Originally SAM was trained with natural images, however, since its publication multiple different domain-specific fine-tuned models have been published. In this workshop, we will focus on using SAM using Python for different segmentation tasks and learn about different parameters to fine-tune results. Participation will be limited. The format will be a combination of tutorials and hands-on exercises.

Time:
Wednesday, February 28, 2024, 1 PM to 5 PM

Place:
Harvard Medical School. Tosteson Medical Education Center (TMEC) 332.

Format:
In-person.

Skill level:
Intermediate

Requirements:
This course will be taught in Python so previous coding experience in Python and a laptop is required.

Scope:
Introduction to SAM
Installation of SAM and relevant packages
Segment everything and visualization
Segment with points prompt
Segment with box prompt
Testing specialized SAM models such as micro-SAM

Cost:
Free of charge.

Applying:
Please apply for a seat. You will receive an automated email confirmation and you can edit your application after submitting.

Deadline for application:
11:59 pm on Wednesday, February 7, 2024.

Selection:
Mostly based on your short motivation. In case of oversubscription, the available spaces will be distributed evenly between groups. You will be informed by email about your application.

Looking forward to see you there, Ranit Karmakar, Antoine Ruzette, Federico Gasparoli, and Simon F. Nørrelykke