InvertNet: a new paradigm for digital access to invertebrate collections

Abstract

InvertNet, one of the three Thematic Collection Networks (TCNs) funded in the first round of the U.S. National Science Foundation's Advancing Digitization of Biological Collections (ADBC) program, is tasked with providing digital access to ~60 million specimens housed in 22 arthropod (primarily insect) collections at institutions distributed throughout the upper midwestern USA. The traditional workflow for insect collection digitization involves manually keying information from specimen labels into a database and attaching a unique identifier label to each specimen. This remains the dominant paradigm, despite some recent attempts to automate various steps in the process using more advanced technologies. InvertNet aims to develop improved semi-automated, high-throughput workflows for digitizing and providing access to invertebrate collections that balance the need for speed and cost-effectiveness with long-term preservation of specimens and accuracy of data capture. The proposed workflows build on recent methods for digitizing and providing access to high-quality images of multiple specimens (e.g., entire drawers of pinned insects) simultaneously. Limitations of previous approaches are discussed and possible solutions are proposed that incorporate advanced imaging and 3-D reconstruction technologies. InvertNet couples efficient digitization workflows with a highly robust network infrastructure capable of managing massive amounts of image data and related metadata and delivering high-quality images, including interactive 3-D reconstructions in real time via the Internet.

Keywords: Collection digitization; collection database; image processing.

Figure 1.

A set of three-dram vials scanned using a color flatbed scanner showing the front (left) and back (right) of the same set of vials. Note that the position of empty spacer vials (e.g., sixth from top in middle column) is the same, but inverted, in the two images because the vial racks are flipped vertically between scans. This relatively quick and inexpensive procedure exposes at least some label data for subsequent capture and reveals the general condition of specimens.

Figure 2.

Current HUBzero-based InvertNet homepage showing top menu bar with content areas accessible to registered users.

Figure 3.

Current version of InvertNet’s Medici multimedia semantic content management system interface, accessible from InvertNet digital collections tab on homepage, showing taxonomic tree, drag and drop file upload space, and zoomable user interface for viewing gigapixel images.

Figure 4.

Image of multiple pinned insect specimens in unit tray (left) and same specimens segmented into separate files (right) using customized ImageJ image processing protocol.