A grant has been awarded to the University of Michigan under the direction of Dr. Inigo Granzow-de la Cerda to use a high throughput method for capturing data from herbarium labels from 100,000 plant specimens in the University of Michigan Herbarium (MICH) that were collected from Mexico and Central America. The collection at MICH is one of the largest in North America for plants from this region. The project will expand upon an earlier project and add to an online catalog for these specimens that can be used by researchers in plant taxonomy and conservation biology worldwide. A key innovation of the work is to use software called HERBIS, which uses optical character recognition (OCR) and natural language processing to automate the extraction of data printed on existing herbarium labels and compile the information into a database. This is one of the first projects to employ HERBIS on a large scale, and a module will be developed for the NSF-sponsored collection software called SPECIFY that will be made available to the community at large. The project leaders at Michigan will collaborate with researchers in Mexico at CONABIO, the federal institution of Mexico charged with studying biodiversity in that country. The interdisciplinary project will employ students in computer science and information science. Undergraduate students in the US and Mexico will be trained in database design and data entry, image capture and processing of textual data, and herbarium management. The project will enhance the biodiversity data for Mexico, in which more than 10% of the world's flowering plant species occur.
