Object detection (dangerous wildlife detection)

• Main Author: Su, Te
• Other Authors: Lu Shijian
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Computer science and engineering
• Online Access: https://hdl.handle.net/10356/166099
• Institution: Nanyang Technological University

Wild animal attacks have cost hundreds of thousands human lives each year worldwide. With the rapid advancement of object detection algorithms, many practical use cases of object detection have been developed for the betterment of our lives. Therefore, this project proposes a real time animal detection model to detect and provide early warnings when encountering a dangerous animal in the wild. In this project, we will determine a suitable model by evaluating 2 types of object detection algorithms, One-stage detector algorithm YOLOv8, and Two-stage detector algorithm Faster R-CNN based on the mean average precision (mAP) score and inference speed. The chosen model will then be integrated to a simple program to detect dangerous animals in real time.