Crowd Anomaly Detection and Performance Evaluation using YOLO Algorithms

In recent years, the YOLO object detection algorithm has gained significant attention due to its superior accuracy and efficiency compared to counter algorithms. This research paper particularly aims to test the YOLOv5 algorithm on various images to detect various objects and compare the same with the YOLOv3 model. This paper also discusses various parameters based on which the algorithm’s efficiency is determined. Our experimental results show that YOLOv5 achieves higher accuracy with low processing speed. The highest accuracy attained by YOLOv5 is 0.99 with F1 score of 0.9. The paper has ability to provide valuable information to the future researchers.