Wildlife-human and wildlife-vehicle encounters often result in injuries and sometimes fatalities. Thereby, this research aims to mitigate the negative impacts of these encounters in a way that makes the environment safer for both humans and animals. The proposed detection system is activated when an object approaches its field of vision, by the use of deep learning techniques, automated object recognition is achieved. For training, we use a labeled dataset from the British Columbia Ministry of Transportation and Infrastructure’s (BCMOTI) wildlife program, and the Snapshot Wisconsin dataset as well. By using Convolutional Neural Network (CNN) architectures, we can train a system capable of filtering images from these datasets and identifying its objects automatically. Our system achieved 99.8% accuracy in indicating an object being animal or human, and 97.6% accuracy in identifying animal species.
CITATION STYLE
Ibraheam, M., Gebali, F., Li, K. F., & Sielecki, L. (2020). Animal Species Recognition Using Deep Learning. In Advances in Intelligent Systems and Computing (Vol. 1151 AISC, pp. 523–532). Springer. https://doi.org/10.1007/978-3-030-44041-1_47
Mendeley helps you to discover research relevant for your work.