Use Case: Field Museum — Whole-Drawer Imaging and AI Workflow at Scale

Organization: Field Museum

Champion: Dr. Tatiana Sepulveda, Collections Manager; Bruno A. S. de Medeiros, Negaunee Assistant Curator of Pollinating Insects;

Products Used: Magnify2 imaging system with backlight and telecentric lenses.

Collection Size: 7 million specimens

 


Elizabeth Postema at the GIGAmacro whole drawer imaging station.

 

Background

The Field Museum’s Insect Collection in Chicago, led by Dr. Bruno de Medeiros (Curator) and Dr. Tatiana Sepulved (Collection Manager), is pioneering a high-throughput imaging workflow to study insect morphology, color evolution, and biodiversity patterns. The development of this workflow is led by Dr. Elizabeth Postema (Postdoctoral Researcher). Using a GIGAmacro Magnify² robotic imaging system, the team has captured over 130 drawers—about 30–40,000 specimens—in the first phase alone. Their goal is to digitize the museum’s entire 7-million-specimen collection, linking detailed imagery to taxonomy, barcodes, and research databases.

 

Challenge

The team faced multiple challenges common to large natural history collections:

  • Balancing scale and detail: Traditional one-by-one imaging was too slow for millions of specimens but essential for detailed morphology studies.
  • Improving visibility: Researchers and curators needed to browse drawers remotely for research or loan requests—something ad-hoc phone photos couldn’t deliver.
  • Data pipeline bottlenecks: Post-processing thousands of images into usable specimen data was time-consuming and error-prone.
  • Operational reliability: A system usable by staff, interns, and volunteers was critical to maintain consistent throughput over long periods.

The ultimate challenge was not just imaging specimens—it was building a scalable, maintainable workflow for years of daily digitization and research use.

 

Solution

The Field Museum implemented a whole-drawer imaging workflow using the GIGAmacro Magnify² platform with automated focus stacking and stitched imaging.

  • Comprehensive coverage: Each drawer is photographed at high resolution, with smaller specimens imaged in grids using template holders.
  • Parallel capture: While one drawer runs, another is staged—keeping the system continuously productive.
  • Automated processing: Post-processing software detects and crops specimens, extracts outlines and measurements, and maps images back to physical barcodes.
  • Volunteer-ready interface: The “press-go” capture workflow enables easy training of interns and volunteers, converting available help into real imaging capacity.
  • Barcode integration: Every specimen now receives a Field Museum barcode, linking dorsal images and metadata directly to collection records.
  • AI Pipeline: The resulting extreme resolution imagery and data feeds directly in an AI pipeline for image segmentation and processing using custom scripting and integration with Roboflow AI tools.

This new process allows the team to capture data faster, with greater consistency and less manual intervention.

 

Preparing a whole drawer capture on the system.

Examples of the AI segmentation of the full resolution imagery during post processing.

 

Results & Impact

The GIGAmacro system transformed imaging into a research-enabling, high-throughput process rather than a tedious chore.

  • Throughput: In a relatively short amount of time, over 130 drawers and tens of thousands of specimens captured across projects, with thousands more in progress.
  • Research acceleration: Whole-drawer imagery made large-scale studies—like beetle color macroevolution—possible within the timeframe of a PhD or postdoc.
  • Operational efficiency:
    • Imaging tasks run in parallel; staff prepare the next drawer while another captures automatically.
    • Non-specialists operate the system confidently after short training sessions.
  • Collections accessibility:
    • High-resolution drawer images allow curators and external researchers to “visit” drawers remotely.
    • Once fully linked to barcodes, researchers will browse individual specimens online and select them for sequencing or loans without handling them in person.
  • Longevity: The team estimates a long term 10–15-year timeline to create first-pass dorsal images and barcodes for all 7 million specimens—an achievable target for the first time in the museum’s history.

 

Representative Quotes

“Any project that needs tens or hundreds of thousands of dorsal images becomes possible within a postdoc time frame. Without whole-drawer imaging, we’d have to rethink the science.”

— Dr. Bruno de Medeiros, Curator

“Automated stacking and stitching—pressing ‘Go’ and letting it run—has been a game-changer. I can stage the next drawer while the current one finishes.”

— Dr. Elizabeth Postema, Postdoctoral Researcher

“Because it’s so reliable, interns and volunteers can run it easily, turning extra hands into measurable throughput.”

— Dr. Bruno de Medeiros

 

Published Data and Resources

The Field Museum is preparing to publish its imaging workflow and dataset publicly, starting with tiger beetles as a pilot group. The team aims to integrate these high-resolution images with genetic, geographic, and ecological data—linking phenotype and environment on an unprecedented scale.

Recently published articles:

https://doi.org/10.32942/X2QW84

https://roboflow.com/case-studies/field-museum

https://academic.oup.com/beheco/article/34/1/1/6618784

In the near future:

  • Expansion to other insect families and taxa
  • Methods publication and data-sharing frameworks

Example of the post-processing steps in the paper Drawer Dissect.

 

GIGAmacro’s Role

GIGAmacro worked closely with the Field Museum team to establish a reliable, high-speed imaging process that supports both research and digitization goals. The collaboration focused on:

  • Supporting queue-based project management and automated processing
  • Incorporating user feedback into new hardware and software refinements
  • Online training and interface refinements for the multi-user team.

“What’s amazing about GIGAmacro is the customer support. If something weird happens, I can reach out directly and get help right away.”

— Dr. Elizabeth Postema