Amphibians such as spotted salamanders pose challenges to scientists who would like to physically tag them for purposes of tracking. Fortunately, there has been a great deal of work in recent years on non-invasive tracking procedures, including photographic capture. The challenge, then, is to identify individuals from digital images and to do so computationally. In this talk I will discuss the challenges of extracting useful data from images of spotted salamanders. I will describe and compare two algorithms for identifying individuals and will discuss why "obvious" approaches that mimic the spot counting used by humans do not perform well.
Bio: Andrea Danyluk is a professor in the department of Computer Science at Williams College. She received her PhD from Columbia University in 1992. Her research interests are in machine learning.