Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{201702,
        author = {Jignesh Chaudhari and Shreeyash Pawar and Dimpal Phalak and Mrs. Purva D Thakare},
        title = {Smart Video Activity Detection System},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {12},
        pages = {4584-4590},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=201702},
        abstract = {An incident occurred in our college in which there was an occurrence of a knife-based threat. This necessitated the development of a Smart Video Activity Detection System that could detect threat situations in real-time. It is based on YOLOv8 Nano model which was trained on a custom dataset containing 4,048 images sourced from various sources like Roboflow, Google Images, Kaggle, and many other different sources. The dataset contains annotated images totaling to 4048 with threat classes/ categories being knife, gun, alcohol, smoking, violence and grenades. We also incorporated YOLOv8 pose estimation to know body posture and the relation of the object with others for better understanding of what activity is going on. The training and testing of the model was done in local PC and got mean Average Precision score of 0.749, precision of 0.817 and recall of 0.734. Our interface is simple and built in Python using Tkinter library supporting real-time video from webcam as well as video file. In case of any detection, a message gets triggered on our email account as well as WhatsApp and a picture is saved with a sound alarm and logging of daily CSV. Our testing confirmed that this system can work effectively in real-world surveillance scenarios like college campuses which will help to prevent a threat.},
        keywords = {YOLOv8, YOLOv8 Nano, object detection, pose estimation, real-time surveillance, weapon detection, violence detection, threat detection, activity recognition, deep learning, computer vision, multithreading, alert system, computer vision, Tkinter, WhatsApp alert.},
        month = {May},
        }