Enhancing Object Detection and Tracking from Surveillance Video Camera Using YOLOv8

In the rapidly evolving landscape of object tracking and detection techniques, video surveillance systems have seen significant advancements, boosting their ability to discern threats and adversarial activities. However, as technology progresses, malicious actors continually adapt their strategies to evade detection. Adversarial attacks, along with incidents like fire outbreaks, violent actions, intrusions, and video tampering, pose substantial risks to the integrity and effectiveness of these systems. Manipulating images and videos can lead to compromised tracking results, either through alteration or deletion of key frames. To address these evolving challenges, researchers and developers are tirelessly working to create more flexible, robust, and resilient object tracking and detection systems. While video object detection is paramount for in-depth scene exploration, it has remained relatively underexplored due to the scarcity of labelled video datasets. The YOLOv8 algorithm harnesses the strengths of the YOLOv8 architecture to elevate object detection performance. This study focuses on real-time analysis of surveillance camera-generated video data, presenting an automated detection approach employing smart networks and algorithms.