Abstract :

The Rapid Increase In Video Data Has Necessitated Smart Surveillance In The Contemporary Security Systems Because It Is Their Requirement To Detect Anomalies In Real-time. In This Survey Paper, Recent Deep Learning Methods Have Been Discussed, Including The Particular Attention To Convolutional Neural Networks (CNNs) And Their Variants Generated And Applied In The Virtual Test Anomaly Detection. The Paper Shows That A Variety Of Background Subtraction, CNN-based Feature Extraction, Autoencoders, Transformer Models, And Hybrid Deep Architectures Are Being Utilized To Detect Abnormal Events With High Precision In An Automatic Fashion. Along Those Lines, The Available Studies Indicate A High Advancement In Detecting Suspicious Actions, Enhancing The Rate Of Detection, And Lowering The Role Of Human Intervention. Nevertheless, There Are Still Largescale Video Processing, Occlusions, Complicated Settings And Generalization Issues. The Survey Covers The State-of-the-art Practices, Their Weaknesses And Limitations, Along With The Most Important Research Opportunities In Order To Create Smarter, More Reliable And Live Surveillance Systems.

Published:

04-5-2026

Issue:

Vol. 26 No. 5 (2026)

Page Nos:

110-119

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite