Scientific Imaging and Machine Learning Workshop

Came across this workshop in my feed. Thought maybe interesting to many CV folks here. And it is in Puerto Rico!

Deadline in 2 days though.

Reminder: Deadline to apply to the Scientific Imaging and Machine Learning Workshop ( Scientific Imaging and Machine Learning Workshop | DSI ) is this Friday Oct 30, 2026!

The 4-day workshop will take place in San Juan, Puerto Rico, March 15 - 18, 2027. The workshop will cover hotel accommodation. Limited travel scholarships are available. Application deadline is October 30, 2026.

Workshop Overview

Imaging workflows build around a core set of computer vision tasks, such as classification and segmentation. While considered mature in computer vision, off-the-shelf models often struggle to generalize to scientific data collected by different instruments under different environments. Too often, solutions are developed within individual disciplines, even when breakthroughs in one field could directly benefit another. For example, U-Net, originally developed for biomedical image segmentation, has now been widely used in almost all scientific imaging fields.

This workshop brings together computer vision and machine learning method developers with domain scientists who work with image-based data across fields like Earth sciences, biomedical imaging, and astronomical observations. This workshop aims to enhance cross-domain and method transferability awareness and identify high-impact, unsolved scientific imaging problems.

Format

Instead of coming to present a poster or a talk, there are two ways to participate:

  • I have an imaging problem I want to solve: What is the biggest imaging bottleneck holding back your research? Bring it here, find out whether scientists in other domains face similar technical obstacles, and learn from computer vision experts and fellow domain scientists alike.
  • I have an imaging method I want to share: If you have developed or adapted a computer vision (or scientific imaging) method to potentially solve a domain-specific imaging challenge, share it with other domain scientists and computer vision experts. This is an ideal venue to extend the impact of your work beyond your own domain, and to learn what remains the unsolved imaging problems.

You will be matched into small groups centered on shared challenges, e.g. domain shift group, memory efficient group. Members in the same group will work together for 1.5 days. Instead of developing a single joint project, you will work on the same technical bottleneck but focus on your own application.

Application

The workshop is open to graduate students, postdocs, early-career researchers within 5 years of their highest-degree, such as faculty or research scientist. All applicants should complete the application form and submit the following two application materials no later by October 30, 2026:

  • A CV.
  • A Statement of Interest, no longer than 2 pages but ideally within 1 page, including figures and tables, excluding references.

Participants will come from a range of fields working with image-based data, including but not limited to, Earth and geophysical sciences (satellite, radar, seismic imagery, and other remote sensing data), biomedical (CT, X-ray, MRI, microscopy, endoscopy, and other biomedical images), astronomical observations, and robotics and autonomous driving.