DASS with Dr. Allison Yi Hsiang
Welcome to the DEEP Advanced Seminar Series (DASS) with Dr. Allison Yi Hsiang. Always at 14.00-15.00 in N399, followed by a fika in the lunchroom.
Applications of deep learning and computer vision in microfossils
Microfossils such as planktonic foraminifera represent an important source of information on marine environments, climate patterns, and geosphere-biosphere dynamics. Traditional methods of collecting data from marine microfossils involve time-consuming manual labor (e.g., picking and measuring objects ~100-200 microns in size) and specialized taxonomic expertise. This limits the size and scope of datasets that can be collected in a reasonable amount of time — which, in turn, limits the types of questions we can ask and answer. In this talk, I will discuss the application of automated methods using deep vision and computer vision to the study of planktonic foraminifera. Using these techniques, many laborious tasks can be automated, such as specimen measuring and identification. By pooling together the expertise of many human taxonomic experts, models with accuracies that surpass any individual human expert can be trained and deployed to identify thousands of individuals in only a few minutes. Using these technologies responsibly — e.g., maintaining careful annotation practices to minimize error — allows for the rapid generation of high-quality datasets that allow us to better understand large-scale community evolution and responses to environmental change across time and space.
Last updated: 2026-06-09
Source: Department of Environmental Science,Department of Ecology, Environment and Plant Sciences