A Decentralized Fog Architecture for Video Preprocessing in Cloud-based Video Surveillance as a Service

With the increasing demand for Cloud-based Video Surveillance as a Service (VSaaS), the efficient processing of vast amounts of video data poses significant challenges. The framework leverages Fog computing at the network edge, enabling real-time video analytics, object detection, and event recognition to be performed closer to the data source. By offloading intensive computational tasks from the central cloud to edge fog nodes, the framework reduces bandwidth consumption and minimizes processing latency. By integrating the state-of-the-art machine learning algorithms, automated video preprocessing with higher accuracy can be achieved. Moreover, the decentralized fog architecture enhances system responsiveness and ensures data privacy by processing sensitive information locally. The primary objective of this preprocessing is to reduce the volume of data transmitted to the cloud while preserving critical information for subsequent analysis. By leveraging fog computing, background subtraction, and intelligent video compression, a substantial reduction is achieved in data transmission, leading to more cost-effective and responsive video surveillance systems.