George Ku

Publication Date

Spring 2025

Degree Type

Master's Project

Degree Name

Master of Science in Computer Science (MSCS)

Department

Computer Science

First Advisor

Philip Heller

Second Advisor

Maya deVries

Third Advisor

Saptarshi Sengupta

Keywords

Autonomous reef monitoring structures, computer vision, coral reefs, data augmentation, mask R-CNN, sponges

Abstract

Coral reefs can be primarily found in tropical and sub-tropical regions of our oceans, providing a thriving habitat for millions of species. Marine sponges, which can be frequently found in coral reefs, play a critical role that contributes to the maintenance of these ecosystems, including the recycling of nutrients through water filtration. However, rising ocean temperatures and acidification due to climate change have resulted in the bleaching and death of coral reefs worldwide. In order to preserve these reefs and the sponges that depend on them, scientists have been performing studies on their biodiversity. This includes collecting numerous images of autonomous reef monitoring structures (ARMS) in order to document the species living in an area. Although this method can give scientists a good idea of the biodiversity in a region, the sheer volume of images collected makes it difficult to quickly and accurately analyze entire datasets. To address this challenge, we employed a Mask Regional Convolutional Neural Network (Mask R-CNN) framework to create a model for automating the process of identifying specific sponges. After training a model on previously selected samples of a unique sponge species, we are able to reliably use this model to identify additional instances of the sponges in other images. We then tested how data augmentation on our training sets affects the reliability of these models. These tests demonstrate the ability of Mask R-CNNs to identify sponges and further introduce the possibility of applying similar models to other species and images.

Download

Available for download on Monday, May 25, 2026

Share

COinS