Auto-detection of Southern sea otters (Enhydra lutris nereis) from aerial imaging on the Monterey Peninsula

Margaret Daly, Carlota Pares-Morlans, Mohammed Salman

A project for CS230: Deep Learning, Fall 2021, Stanford University, CA.

Project Tree

The main file for this project is ./src/YOLOv5_SeaOtter_OD.ipynb. It calls the necessary files from the ./src/yolov5-master folder and trains the modified YOLOv5 model for small object detection.

The folder ./src/utils includes some functions that have been used for preprocessing and analyzing the dataset, as well as some helpers for visualizing the outputs.

.
|-- ./sample_dataset
|-- ./src
    |-- ./src/utils
        |--./Sea_Otter_Error_Analysis.ipynb
    |-- ./src/YOLOv5_SeaOtter_OD.ipynb
    `-- ./src/yolov5-master