StripeSpotter Scanning System helps Scientists identify Animals in the Wild
Researchers at Princeton and the University of Illinois have recently released a fascinating research paper that details a scanning system for recognizing animals through their unique markings. The GNU licensed Stripespotter system is an open source project designed to automate animal recognition for professionals studying them in the wild by recognizing patterns of pixels in digital photographs. Traditionally scientists have had to use tracking mechanisms or excrement to identify the animals they study, but now they can simply take a picture of an animal and query a database to find biometric information or create new entries.
The researchers compiled a dataset based on two species of Zebra in Kenya to augment the efforts of local field ecologists. The algorithm they developed uses images taken by cameras with a high resolution of 8 megapixels or more, suggesting that many smartphones have the hardware capabilities to utilize the system. Currently a researcher takes a picture with a digital camera, and then uploads the photo to their computer where they highlight a section of the photograph. The highlighted portion is then compared to the entries in the database to identify the animal. Field tests have shown that Stripespotter can be applied to animals with a variety of different morphological characteristics such stripes or spots by tests conducted on both Zebras and Giraffes. Soon we may see the use of this technology expanded to other exotic animals such as Tigers, Cheetahs and Clownfish.
The researchers have provided open access to their dataset to kick start the efforts of developers and field professionals who are interested in becoming first movers in this new area of scanning technology. While the current system seems simple and easy to use, we’re looking forward to some juicy mobile apps being created as a result of this research. We’re already participating in this movement through our all new Scanimal feature that enables “natural” barcode scanning. We’ll be keeping our eye out for more on this emerging area of research, and reporting the latest and greatest breakthroughs here on the Scandit blog.