1Assistant Professor, Department of Information Technology, Manakula Vinayagar Institute of Technology, Puducherry, India. Email: praba0885it@gmail.com
2,3,4,5UG Student, Department of Information Technology, Manakula Vinayagar Institute of Technology, Puducherry, India.
Background: Global security still presents difficulties demanding improved techniques for real-time threat identification. Conventional object recognition models similar to RCNN battle with accuracy and speed, thereby restricting their surveillance efficacy functions. VGG19, a deep learning based convolutional neural network, tackles these problems with their remaining connections, so enabling more thorough network training free from vanishes gradients here. This architecture improves quality, extracting, raising precision while preserving computing effectiveness.
Methodology: Unlike RCNN, VGG19 handles pictures quicker, therefore fitting for real-time investigation. Its capacity to identify risks with regard to more accuracy guarantees timely reactions, lowering possible security dangers.
Conclusion: Using VGG19, surveillance systems can attain improved threat recognition, strengthening security in several surroundings. As such Security threats change and using effective Models such as VGG19 become absolutely crucial for strong and proactive security solutions.
Key terms: threat, real-time surveillance Detection, RNN, VGG19, Deep Learning, Residual Connections; Object Detection.
How to cite this article: Prabavadhi J, Sindhuja R, Subha Sri R, Roja N, Vasanthy D. Optimizing Security Surveillance with VGG19: A Deep Learning Approach to Real-Time Threat Detection. Int J Drug Deliv Technol. 2026;16(13s): 872-878. DOI: 10.25258/ijddt.16.13s.96.
Source of support: Nil.
Conflict of interest: None