deep learning for octopus ID/detection. i explore solutions for practical computer vision + lightweight ml workflows to assist with: (1) species identification, and (2) object tracking. octomatic is a tool to assist with octopus identification that is trained on expert-labeled datasets to provide more accuracy than internet trained classifiers.

tracking movement. i am investigating novel methods of tagging + tracking octopuses in south florida. currently, i am exploring visible implanted elastomer for tag-recapture, as well as novel bioadhesives to attach fine-scale trackers without sutures or implants. ideally, these methods will allow for better understanding of octopus movement, how environmental factors influence their movements, and prove the viability of a non-invasive tagging methodology for future octopus research.

sympatric octopus interactions. using annotated video and ethograms, i am exploring how conspecifics and interspecific interact. blue heron bridge sits in a shallow water lagoon home to multiple octopus species, and i am investigating how these species influence each other and their environment.