Create your own conference schedule! Click here for full instructions

Abstract Detail

Bioinformatic and Biometric Methods in Plant Morphology

Punyasena, Surangi [1], Mander, Luke [2], Li, Mao [3], Mio, Washington [3], Fowlkes, Charless [4].

Classifying grass pollen using high-resolution imaging and the quantitative analysis of surface texture.

The taxonomic identification of pollen and spores uses inherently qualitative measures of morphology. As a result, identifications are restricted to categories that can be reliably classified by multiple analysts. A consequence of this conservative approach is the coarse taxonomic resolution of the pollen and spore record. Grass pollen represents an archetypal example; it is not routinely identified below the family level. To address this issue, we explored high-resolution imaging methods, including scanning electron microscopy (SEM) and super-resolution structured illumination microscopy (SR-SIM), to digitally capture the surface ornamentation of grass pollen and developed quantitative morphometric methods to characterize and classify grains. In doing so, we produce a means of quantifying morphological features that are traditionally described qualitatively. We used SEM to image 240 specimens of pollen from 12 species within the grass family (Poaceae). We demonstrate that similar features can be observed using SR-SIM. We classified these species by developing features that quantify the size and density of sculptural elements on the surface of each pollen grain and measure the complexity of the ornamentation that they form. Our experiments using these features yielded a classification accuracy of 77.5%. We compared our approach to a texture descriptor based on modeling the statistical distribution of brightness values in local image patches. This approach is common in computer vision and yielded a classification accuracy of 85.8%. However, the quantitative features we developed in this paper yield the additional benefit of directly relating to biologically meaningful features of grass pollen morphology. We anticipate that their use in future work will allow palynological experts to better interpret unsupervised classification results in the fossil record, which currently is not possible with standard machine-based approaches.

Broader Impacts:

Log in to add this item to your schedule

1 - University of Illinois, 505 S Goodwin Ave, 139 Morrill Hall, Urbana, IL, 61801, USA
2 - Plymouth University, Earth Sciences, School of Geography, Earth and Environmental Sciences, Room 121, Fitzroy, Drake Circus, Plymouth, Devon, PL4 8AA, UK
3 - Florida State University, Mathematics, Tallahassee, FL, 32306-4510, USA
4 - University of California Irvine, Computer Science, Irvine, CA, 92697, USA

none specified

Presentation Type: Symposium or Colloquium Presentation
Session: C2
Location: Prince of Wales/Riverside Hilton
Date: Monday, July 29th, 2013
Time: 3:15 PM
Number: C2007
Abstract ID:646
Candidate for Awards:None

Copyright 2000-2012, Botanical Society of America. All rights reserved