Based on Machine Learning: Differentiating between Sea Otters and River Otters

Machine learning has become a critical tool in biological studies for both prediction and discovery. This paper specifically focuses on distinguishing between sea otters and their similar species, the river otter, using Edge Impulse to develop models deployable on mobile devices and ESP32 platforms. Addressing a gap in existing research, we collected and appropri-ately processed our own data, ultimately assembling a small but robust dataset. This dataset was divided into training and test sets for the purposes of this study. We utilized the MobileNetV2 neural network for training, successfully developing two models with accuracies of 91.67 % and 83.33 %, respectively, each with different computational overheads. This research not only aids in accurately identifying sea otters, an endangered species, but also contributes to wildlife conservation efforts, demonstrating substantial practical applications.