The Californian CESU Berkeley is excited to announce a new publication within the Aquatic Conservation Journal. “On‐Land Pinniped Classification of Multiple Species and Demographic Classes on Multiple Substrates Using Deep Learning and Aerial Imagery” co-authored by Silas Santini, Sarah Codde, Elizabeth M. Jaime, Alan Jain, Esteban Valenzuela, and Benjamin H. Becker, explores how deep learning can enhance pinniped demographic monitoring.
As stated in the article, monitoring pinnipeds in California is a large endeavor, and can prove significantly challenging. Current methods include aerial surveys and ground counts, both of which are time consuming and susceptible to human error. Deep learning in combination with high quality aerial imagery can reduce these drawbacks. Reflecting on the study’s potential impact, lead author Silas Santini explains, “Our promising results show that fine tuning machine learning models can produce fast and inexpensive identification of pinniped species by age and/or sex while on multiple substrates. If our workflow were to be adopted, seal counts could be more accurate, quicker, and safer. Accurate counts are important for the protection and management of these species, and others. After all, our methods came from a study on seabirds and could be adapted for other animals”.
To learn more, read the article here.