International Display Workshops General Incorporated Association

[ISTp1-2] Development of Goose Pool Stay Duration Evaluation System Based on Deep Learning Technology

*Shuo Ching Sun1, Zi-Heng Jian1, Yao-Chuan Tsai1 (1. National Chung Hsing University (Taiwan))

YOLO v4, deep learning, goose, tracking

https://doi.org/10.36463/idw.2023.1494

In this study, we developed a geese duration evaluation system using YOLO detection and tracking algorithms. The experimental result indicated that the trained detection model achieves excellent precision and recall rates of 94.6% and 84.1%. Also, the tracking algorithm achieves a tracking accuracy of 87%, enabling timely measurements for farmers