Carnegie Mellon University Research Showcase @ CMU Fine International Conference on Gigapixel Imaging for Science Conferences and Events 11-2010 Remote curation and outreach: examples from the NCSU Insect Museum GigaPan Project Matthew A. Bertone North Carolina State University Insect Museum Andrew R. Deans North Carolina State University Insect Museum Follow this and additional works at: http://repository.cmu.edu/gigapixel Published In Proceedings of the Fine International Conference on Gigapixel Imaging for Science. This Conference Proceeding is brought to you for free and open access by the Conferences and Events at Research Showcase @ CMU. It has been accepted for inclusion in Fine International Conference on Gigapixel Imaging for Science by an authorized administrator of Research Showcase @ CMU. For more information, please contact research-showcase@andrew.cmu.edu.
Remote curation and outreach: examples from the NCSU Insect Museum GigaPan Project Matthew A. Bertone NCSU Insect Museum, North Carolina State University, Raleigh, NC, 27695 matthew.bertone@gmail.com Andrew R. Deans NCSU Insect Museum, North Carolina State University, Raleigh, NC, 27695 andy_deans@ncsu.edu ABSTRACT Insect specimens preserved in publicly accessible, permanent collections are integral data resources for researchers and a valuable tool for engaging the public about biology. Natural history museum policies intended to protect specimens from damage, however, usually prevent people from accessing these insects in meaningful ways. We propose using the GigaPan system as part of a larger collection digitization effort to enable virtual exploration of our specimen holdings through high-quality images of all of our insect drawers (~2,700 in total). We envision at least two outcomes from this process: a) researchers worldwide will be able to remotely identify insect specimens and/or read their associated data labels, and b) nonentomologists will gain a new resource with which to learn about insect diversity, biology, and structure, as well as museum science more generally. Capturing the dialog from these two processes, as annotations, snapshots, emails, and other correspondence, will lead to iterative improvement of our collection, more efficient loan requests, more informed specimen donations, and highly effective outreach experiences. Keywords Insecta, specimens, collection, Canon G11, entomology, outreach INTRODUCTION Formally established in 1952, the NCSU Insect Museum maintains one of the two largest collections of insects in the southeastern United States, while serving as a resource for leading-edge research, innovative training, and inspiring outreach. The Insect Museum provides critical specimen-based information that is actively used to confirm identifications, to understand insect distributions through time (especially the decline of sensitive species and the expansion of exotics), to document the diversity of insect phenotypes, and to elucidate patterns of evolution. Insect specimens are fragile and should not be exposed to adverse environmental conditions or unnecessary handling, including the movement and opening of drawers. As such, and consistent with the policies of most research collections, the Insect Museum is only open by appointment. Browsing our holdings for educational and/or research purposes remains impossible without physical access to the cabinets. We intend to image all 2,700 drawers of pinned specimens housed in the Insect Museum using GigaPan technology and to avail the subsequent images in a way that permits virtual browsing and annotation. We predict that remote curation and outreach will result (minimally) in higher quality and more frequent taxonomic determinations (i.e., identification), more precise loan requests, more informed donations, and a broader reach for educating the public about insects. MATERIALS & METHODS Equipment In this ongoing effort we employ a GigaPan EPIC 100, oriented horizontally on a copy stand and outfitted with a Canon G11 camera. We retrofitted the GigaPan with an A/C adapter (1) and bought a commercial A/C adapter for the Canon. Our lighting needs are satisfied by duel Interfit Super Cool-Lite 9 lights, each with nine 28W compact-fluorescent bulbs. These bulbs produce continuous daylight spectrum (5000-5500K). The complete imaging station is illustrated in Figure 1. Future enhancements may include an X-Y rig for positioning the camera directly above the drawer at every image position. We will also evaluate options for integrating a digital SLR camera into the workflow, depending on the long-term quality and utility of our current system. 1
Settings Camera settings were selected based on those described on the GigaPan website (2), with the white balance set to daylight fluorescent (best balance for the lighting described above) and the field of view (FOV) for the Canon G11 set to 11.5º on the GigaPan unit. The distance of the GigaPan unit (plus camera) was set at about 40 cm above the insect drawer to optimize the quality of the images (at full optical zoom), while reducing curvature. This distance also allows for sharp enough/close enough images of the specimens while keeping the number of photos (n=35) manageable with respect to time and storage capabilities. Imaging Workflow Drawers are placed within the confines of a custom jig on the copy stand, and the lid is removed. To prevent white space from interfering with the camera s ability to focus, a Kodak Tiffen Color Separation Guide (ASIN: B00009R7G9; trimmed to fit inside a unit tray) and printed matter are placed inside empty unit trays (see Figures 2-3). The Epic 100 is then engaged, using the Last Panorama function (unless the image area needs to be modified), and all photos (n=35, usually) are delivered directly onto our PC from the camera. Stitching is initiated using GigaPan stitcher software (version 1.0.0520) either immediately before the drawer is processed, labeled as imaged, and deposited back in the collection, or as a batch job overnight. Stitched sets are stored locally, backed up by external hard drives, until they can be delivered to the GigaPan website or published on our own server. Figure 4 illustrates a typical drawer image, while Figure 5 shows how specialty drawers can be assembled and imaged to show insects by theme (in this case the diversity of the four largest insect orders). RESULTS & DISCUSSION Though we are only in the beginning stages of this project, tangible results are already becoming apparent. As of 28 June 2010 more than 300 drawers have been imaged, each taking approximately 10 minutes to complete (excluding stitching, which varies depending on the computer processing speed). Twenty-one of these drawer images have been exposed through the GigaPan website (e.g., http://gigapan.org/gigapans/49310/), which yielded both expected and unanticipated results: Unsolicited, remote curation happens. Word of our insect drawer images spread quickly amongst insect systematists, and we very quickly received communications that enhanced our holdings. In one incident, a taxonomist at a natural history museum in Ottawa, ON, 837 miles north of the Insect Museum, determined a series of froghopper specimens (Hemiptera: Cercopidae) in an unsorted insects drawer (Figure 8). In another incident, a world bumble bee (Hymenoptera: Apidae: Bombus) expert was able to provide a species name for an undetermined specimen (Figure 6). Data enrichment happens. An unanticipated outcome of this project is the linking of specimen snapshots to panoramas of their locality/habitat (based on label information). Unsolicited, a GigaPan user took a panorama of the cloud forest habitat in Costa Rica where one of our leafhopper specimens was collected (linked through a snapshot) (Figure 7). The practical applications of these data are plentiful, including using the panorama of the habitat to estimate plant diversity related to insect specimens or change in habitat over time. Unsolicited, expert curatorial advice happens. Taxonomists perused our available images, including our entire Fulgoroidea collection, and provided us with corrected determinations and advice regarding the future reorganization of our collection. We successfully reached out and engaged non-experts. Non-entomologists commented on artistic representation (3; Really like the way the labels, pins and beetles all come together from this angle ), made humorous comments about the insect specimens (4; I would call this a Fu Manchu beetle! ), and asked us questions about insect biology (5; Please explain in objective scientific but lay-understandable detail: whether insects can a) suffocate, and/or B) feel pain ). Gigapans enable informed donations. A world expert has contacted us to say she is using our GigaPan images to better understand our current holdings, so that she can then divide up her personal collection between natural history museums more efficiently. She wants to maximize the taxonomic coverage of her donation to our museum. Another result that is perhaps less tangible, or at least difficult to measure using discreet performance metrics, is the buzz that this project has generated amongst our entomologist and museum colleagues. Most experts readily recognize the utility of drawer GigaPans, and the project has triggered several conversations about how to extend their outreach and research potential, as well as their ability to increase institutional awareness (both internally and externally). CONCLUSIONS The GigaPan system appears to be a user-friendly, efficient, and effective way to digitize insect specimens, and we anticipate that this technology will be tested extensively as part of the upcoming national effort to rescue natural history collections from a largely analog environment (6). The GigaPan website serves as a suitable repository for these images for now, given 2
the established infrastructure, tools, and support. We are also exploring more biological relevant image archives (e.g., http://morphbank.net/) that have formal connections to biological ontologies (7), taxonomic name databases (8), specimen data (9), genetic data (10), and more sophisticated annotation capabilities. Future research using the complete GigaPan image set may focus on automated methods for text extraction (e.g., label data), species recognition (i.e., image recognition), crowdsourced digitization, and unsupervised machine learning exercises. FIGURES Figure 1. GigaPan setup for imaging insect drawers. 3
Figures 2-3. White space fillers. 2) Kodak color standard (left) and 3) narrative about the project. 4
Figure 4. GigaPan image of a typical insect drawer filled with pinned plant hopper specimens (Hemiptera: Flatidae); for more details visit the original image: http://gigapan.org/gigapans/52072/. 5
Figure 5. GigaPan image of a specialty drawer illustrating the diversity of the four largest orders of insects (Hymenoptera, Coleoptera, Lepidoptera and Diptera) specialty drawers can be used to display insects in different ways for outreach purposes; for more details visit the original image: http://gigapan.org/gigapans/49310/. 6
Figure 6 (left). This snapshot of a bumble bee demonstrates the utility of GigaPan images for remote determination of specimens. Though one cannot see its labels, the diagnostic features are easily visible. See snapshots for detailed explanation: http://gigapan.org/gigapans/49310/snapshots/147250/, http://gigapan.org/gigapans/49310/snapshots/147253/, http://gigapan.org/gigapans/49310/snapshots/147254/, http://gigapan.org/gigapans/49310/snapshots/147255/. Figure 7 (right). The specimen in this snapshot is likely too small for remote determination through a GigaPan of this resolution, but one can read almost the entire contents of its collecting event label. A fellow GigaPan user read the label and then proceeded (unsolicited) to GigaPan the Costa Rican cloud forest habitat where this specimen was collected: http://www.gigapan.org/gigapans/42163/. Figure 8. A series of froghopper (Hemiptera: Cercopidae) specimens that were determined remotely (unsolicited) as belonging to Microsargane vittata (Fowler) and then re-curated based on that expert annotation. http://gigapan.org/gigapans/41421/snapshots/120403/. 7
ACKNOWLEDGMENTS Funding for this digitization effort was provided by the National Science Foundation (DBI-0847924; A. R. Deans and B. M. Wiegmann, PIs). We thank Bob Blinn, Tanner Stanfield, Kelly Dew, and Lydia Abernethy for their industrious efforts to prepare and image the pinned insect collection. We also thank Mary Jo Knelly (CREATE Lab at the Robotics Institute, Carnegie Mellon University) for engaging us in the early stages of our project and for providing links to resources and advice. REFERENCES 1. Sargent, R., Denning, S. and Bertone, M. (2010) Using an external power adapter with your GigaPan. Available: http://bit.ly/gigapanacadapter Accessed 16 June 2010, 15:42. 2. GigaPan Systems. (2010) User Guide for Gigapan Epic and Epic 100. Available: http://gigapansystems.com/setting-upthe-gigapan-imager.html Accessed 22 June 2010 21:18. 3. Insect Drawer (scarabs) Take 2 GigaPan. http://gigapan.org/conversations/119341/ 4. Insect Drawer (scarabs) Take 2 GigaPan. http://gigapan.org/conversations/117944/ 5. Flatidae - drawer 4 GigaPan. http://gigapan.org/conversations/147239/ 6. NESCent Digitization Workshops Participants (2010) A Strategic Plan for Establishing a Network Integrated Collections Alliance. Available: http://digbiocol.wordpress.com/ Accessed 16 June 2010, 15:42. 7. Yoder M. J., Seltmann K. C., I. Mikó, M. A. Bertone, and A. R. Deans. (2010) Hymenoptera Anatomy Ontology: http://purl.oclc.org/net/hym-ontology. 8. Integrated Taxonomic Information System (ITIS) http://www.itis.gov Accessed 22 June 2010 21:18. 9. Global Biodiversity Information Facility (GBIF) http://www.gbif.org/ Accessed 22 June 2010 21:18. 10. Benson D.A., Karsch-Mizrachi I., Lipman D. J., Ostell J., Wheeler D. L. (2008) GenBank. Nucleic Acids Research 2008 Jan; 36 (Database issue): D25-30. Epub 2007 Dec 11. 8