Microplastics, Machine-Learning, and Multimodal Optical Imaging

Dr. Greg Madjeski

CEO Parverio Inc.,

Visiting Scientist, Biomedical Engineering

University of Rochester

Bausch & Lomb Room 109 or Remote Zoom

ABSTRACT: TMicroplastic debris (1-5000 microns) is everywhere around us. As the demand for plastic production grows (420 million tons annually in 2015), so does microplastic pollution. We are increasingly exposed to small plastic particulates that have the potential to absorb harmful substances from the environment and then cross tissue barriers in the gut. We need to replace labor-intensive and time-consuming processes to enumerate and categorize many thousands of microscopic debris particles and accurately assess the risks of microplastic pollution. Using silicon nanomembrane filtration technology developed at the University of Rochester, we have used a combination of imaging and machine-learning techniques to rapidly prepare and analyze microparticulates from water samples, reducing sample turnaround time from 40+ hours to 20 minutes. An overview of imaging approaches used to tackle this problem will be discussed, including fluorescence, Raman, FTIR, and birefringence, as well as computer vision technique.

BIO: An expert in biosensing, Greg has degrees from the Rochester Institute of Technology (Microelectronic Engineering) and the University of Rochester (Biomedical Engineering), having also studied at the University of Nottingham as a Whitaker Fellow. During his ostdoctoral work, he developed new methods to detect microplastic pollution (Steadman Postdoc Prize, 2nd place) and has recently formed a startup (Parverio Inc, Falling Walls Venture Finalist 2021) to commercialize these methods. Parverio’s mission is to combat microplastic contamination in our environment, food, and drink by providing new tools and information to improve human health outcomes with microparticulate analysis.