Applying Machine Learning to study the social network of Bottlenose Dolphins (Tursiops truncatus) from Bocas del Toro, Panama using photographs of their dorsal fins.
Chavrier, Annalena
Chavrier, Annalena
Citations
Altmetric:
License
License
Abstract
Bottlenose dolphins (Tursiops truncatus) live in complex social systems where individuals often move between groups. Because of this, it can be difficult to understand how the form groups and long-term associations. Additionally, because they are small, they cannot be easily tagged with instruments to learn about their whereabouts and their social patterns. For this reason, marine mammal biologists use photo-identification a non-invasive method that uses the unique shapes, scars, and markings on dolphin dorsal fins to tell individuals apart. These marks have been shown to be relatively stable over time but also need to be updated as dolphins gain new scars as they get older. For long-term studies, keeping track of these individual changes over thousands of photographs can be difficult. New machine learning computational tools, however, can facilitate cataloguing and matching dorsal fins more efficiently. The goal of this study is to use these tools to catalogue and identify individuals from a small population of bottlenose dolphins in the Bocas del Toro Archipelago, which the Ondas Lab has been studying since 2004. When completed, this 20-year catalog will support social network analysis for this population, helping us understand how long-term events, such as the COVID-19 lockdowns, hurricanes, and increased development in the area, affect social dynamics. Currently, the catalog includes photos collected across 14 days in 2024, 12 days in 2022, and 20 days in 2018, identifying a large number of individual dolphins. Long-term photo-identification datasets like this are an important tool for understanding the social structure of dolphin populations. Maintaining and expanding these datasets supports ongoing research on how dolphin social relationships may influence communication and behavior.
Description
Date
1/1/2026
Student Status
Senior (Graduating in 2026)
Journal Title
Journal ISSN
Volume Title
Type of presentation
Poster
Collections
Research Projects
Organizational Units
Journal Issue
Citation
DOI
Advisor(s)
Department
Program/Major
Zoology
College/School
College of Arts and Sciences
Organization
Research Category
Life Science