Challenge:
Red abalone (Haliotis rufescens) is a critically endangered species. Determining sex for breeding programs traditionally requires invasive methods that stress animals and increase operational costs for aquaculture producers.
Approach:
Developed deep learning classification models to determine the sex of red abalone from ultrasound images of reproductive tissue, providing a non-invasive alternative that reduces stress on animals and costs for producers.
Result:
Published in Frontiers in Artificial Intelligence (2026). Lead author. Demonstrated that convolutional neural networks can reliably classify biological reproductive tissue from ultrasound imagery, replacing invasive methods that stress animals and increase operational costs. When deployed on an edge AI device, this approach scales to facility-wide phenotyping without requiring trained technicians at every station. Opens a pathway for non-invasive, high-throughput phenotyping across aquaculture species.
