You are here

1st Place, 2011

Zebraprinting

Mayank Lahiri, Computer Science

Identifying individual animals in the wild is a fundamental step in ecological analysis, used for everything from population size estimation to social network analysis. Animals are usually identified manually from photographs taken in the wild, which is a painstaking, error-prone process. To speed up this process and reduce identification errors, I have developed a computer-aided identification algorithm for wild zebras, similar to a fingerprint or face recognition system for humans. Using photographs captured in the wild, a conservation worker can determine if a zebra has been sighted before, and bring up various types of information on prior sightings. The program is freely available to anyone, thanks to a UIC Provost's Award, and is currently in use at two field sites in Kenya: the Ol'Pejeta Conservancy and the Mpala Research Center.

The image in the submission was captured by me in December 2010 on the Ol'Pejeta Conservancy in Laikipia province, Kenya. It shows a 'zebraprint', or how my computer program perceives the stripes of a zebra for identification purposes. Although the computer representation of zebra stripes looks similar to human fingerprints, the identification algorithm is more closely related to algorithms that are used to analyze DNA sequences.