Abnormal Activity Detection by Leveraging AI

Abnormal Activity Detection by Leveraging AI

Author Name : Shaik Rasool , Adeeba Fatima , Maliha Mubashira Naaz , Mohammed Mudassir Uddin

ABSTRACT In today's digital age, effective surveillance and anomaly detection systems are crucial. This guide details a systematic approach to implementing YOLOv5 for detecting suspicious activities in images and videos. YOLOv5, known for its accuracy and speed, is ideal for real-time object detection. This begins with setting up essential libraries and dependencies to ensure a smooth workflow. It covers importing and curating datasets, emphasizing YOLO format annotations. Exploratory data analysis, including visualizing samples and analyzing class distribution, ensures data quality. The core involves training the YOLOv5 model and evaluating its performance. Practical applications include real-time and batch inference, making it a valuable security tool.