From: Barbour et al (in review), “Clustering and classification of vertical movement profiles for ecological inference of behavior”

We developed a generalizable classification tool for sea turtle dive profiles, using an image classifier convolutional neural network (CNN) and 5 classes of known dive shapes for sea turtles (Hochscheid 2014).

This tool has been fully trained using these 5 standard dive shapes (A, C, D, E, and F) known for sea turtles in combination with a dataset of dive profiles from migrating leatherback turtles in the South Pacific ocean gyre (see Shillinger et al 2008, 2011 for original data descriptions).

Copyright © 2016 Skynet, Inc. All rights reserved.