Work

Abnormal Human Activity Detection Using Yolov5

Year: 2025
Type: article
Abstract: effective surveillance and anomaly detection systems are crucial in the digital age. For detecting suspiciousactions in photos and videos, this book offers a step-by-step method to implementing cutting-edge object recognitionalgorithms—including YOLOv5, YOLOv6, YOLOv7, and YOLOv8 Cutting-edge its outstanding performance,YOLOv5, renowned for its precision and quickness, is the main emphasis. Ensuring a seamless process, the guidestarts by describing the required libraries and dependencies for the project. It then addresses dataset imports, stressingthe YOLO format for annotations, and dataset preparation for efficient model training. Sample photos with boundingboxes are visualised and class distribution examined using exploratory data analysis (EDA). The book also coversevaluating image size and resolution compatibility with YOLOv5 criteria. The main part emphasises using theYOLOv5 algorithm, including model loading, training, and performance evaluation measures. The trained model'spractical usesstate, such real-time or batch inference for security and surveillance objectives, are also investigated.Although YOLOv5 is praised for its great accuracy, the book advises experimenting with later versions—includingYOLOv6, YOLOv7, and YOLOv8—for better performance
Language: English
Open Access status: diamond
OpenAlex ID: W7218138270
FWCI:
Cites: 10
Cited by:
Related to: 10