Wildlife Detection and Recognition in Digital Images Using YOLOv3: Extended Abstract
Mina Gabriel, Sangwhan Cha, Nushwan Yousif B. Al-Nakash, Daqing Yun · 2020
Recent advances in hardware capability and machine learning techniques enable convenient monitoring of wildlife and their living environments. In this work, we apply Deep Learning (DL) methods to detect and recognize wildlife in digital images and report the experimental results conducted in a commodity workstation. Specifically, YOLOv3 and YOLOv3-Tiny are used to detect and classify several classes of animals based on 9051 digital images and they achieve 75.2% and 68.4% mean average precision, respectively.