MOHASA: A Dynamic Video Synopsis Approach for Consumer-Based Spherical Surveillance Video
S. Priyadharshini, Ansuman Mahapatra
Abstract
Using spherical cameras for video surveillance applications can solve the problems faced by traditional cameras in a multi-camera setup. Video synopsis is an object-based approach for reviewing a long recorded video through an optimization process. From the literature review, it is observed that Aquila Optimizer (AO) algorithm has a powerful global search capability but it can easily be trapped in a local optimum. Hence, an improved multiobjective hybrid AO with Simulated Annealing (SA), MOHASA, is introduced. The most challenging situation of using spherical cameras is the Fear of Missing out (FOMO), i.e., choosing to view one portion forces to miss the other portions of the scene. It is solved by dynamically generating video synopsis using both spatial and temporal summarization. Also, a new frame packing constraint has been integrated into the objective function to better understand spherical scenes. Experimental analysis shows that the proposed framework outperforms the state-of-the-art works.