Deploying Phytoplankton Classification Deep Learning Models on Edge Devices

Harmful algal blooms (HABs) are causing significant damages and losses for fish farmer, and therefore must be regularly monitored. Given recent breakthroughs in deep learning, computer vision algorithms can now be created to automatically detect harmful phytoplankton in water. This research explores how to train these models in a secure manner while still meeting client data privacy requirements, as well as how to deploy these models in the field on edge devices. The successful completion of this research will help the partner organization build a system that can be deployed at fish farms around the world, allowing fish operations to better monitor and manage the impact of HABs.

Intern: 
Jason Deglint
Faculty Supervisor: 
Alexander Wong
Province: 
Ontario
Partner: 
Partner University: 
Discipline: