Audio Surveillance System(ASS): Merging of Acoustic Events Recognition and Classification through Deep Learning

Despite the widespread use of Closed-Circuit TV for security purposes, CCTVs still include several weak points that can affect the quality of the provided surveillance services. Scientists and engineers are taking advantage of the continuous advancements in computer science by utilizing technology to improve acoustic recognition & classification approaches with the most efficient manners but at the cost of higher complexities. In this research, we focused on overcoming missing data paired with the use of existent surveillance approaches by adding acoustic surveillance features. We propose a low-complexity fully optimized algorithm that combines acoustic’s recognition and classification algorithms. The proposed Audio Surveillance System (ASS) algorithm uses acoustic input which is then processed through the algorithm using the employed deep learning approaches to extract irregular patterns and classify them into appropriate categories to be used as a surveillance input.