Integrated Interoperability Based Panoramic Video Synopsis Framework
Palash Yuvraj Ingle, Young‐Gab Kim · 2024
The surge in intelligent video surveillance systems for public safety has brought about technological advancements but poses challenges, including handling vast video data. Issues such as extraction, analysis, storage scarcity, and hyperconnectivity emerge as a consequence. Surveillance videos often feature extended periods of empty frames, underscoring the need for efficient information extraction. Video synopsis has emerged as a key framework for streamlined storage and analysis in surveillance. However, the existing video synopsis methodology faces limitations in dealing with the heterogeneous nature of devices that capture videos in various formats (e.g., drones, Ip cameras, wired cameras) and using diverse software and frameworks. Therefore, constructing a video synopsis in heterogeneous platforms is challenging because it raises many interoperability issues. The existing video synopsis methodology is not applicable to heterogeneous devices because they create a synopsis only for a single-view camera. To address these issues, we propose an interoperability-based design to create a panoramic video synopsis. Initially, videos synchronized from different platforms are converted into a standardized format. Panoramic stitching was then applied to create a panoramic view, which was processed to extract the foreground tubes and background, facilitating the construction of a comprehensive synopsis.