Video Synopsis for Interaction Identification: a Graph-Based Approach

Security cameras have been used for several decades to monitor activities. Video synopsis techniques have been used to condense the large amount of video data captured. Video synopsis uses an object-based approach to optimize the review of lengthy recorded footage. However, this technique often misses critical details and interactions while rearranging tubes (tracks of moving objects). To address this limitation and maintain the integrity of the original video, we suggest using a graph-based video synopsis. This method aims to identify tubes interacting with each other employing an interaction matrix. By analyzing input surveillance video, we determine whether surveillance efficiency. Additionally, our approach, grounded in graph preservation, helps in understanding the intricate connections between various elements. This method strategies for optimal placement of moving objects within the output video preserving the interaction and temporal order.