Developing A Framework For Smart Surveillance System Using Machine Learning
DOI:
https://doi.org/10.64252/e3ygk048Keywords:
Deep Learning, CNN, GRU, InceptionNet, Security, Real-time Monitoring, Computer Vision.Abstract
There have recently been more rapid advances in AI and deep learning. Some of these advances provide intelligent surveillance systems autonomously capable of monitoring and analyzing data in most cases in real time. In this paper, a Smart Surveillance System is now described employing several deep-learning-based smart algorithms, specifically InceptionNet and Gated Recurrent Units (GRUs), for very high accuracy in detecting and classifying suspicious activities. The architecture of the system consists of frontend and backend modules. The backend modules consist of dataset acquisition, splitting, preprocessing, and training using the smart algorithms mentioned above. The system's user interface is provided in the front for registration, login, and data input. The security authorization mechanisms will ensure that only entries from classified authorized personnel can access the system. After authentication, the input data are sent to the training model to generate pertinent insights. The results will be presented to the user; thus, timely decision-making will be reinforced in the security monitoring. It is the adopted integrated system, which promises to offer the most reliable, scalable, and efficient modern surveillance applications in vulnerable settings like public places, transport terminals, and private establishments.