Security Surveillance Video using CNN With SMS Alert

Abstract

In recent times cameras have been mounted in many different locations for security and surveillance purposes. The inspection of the data that is captured using the surveillance system can play  an vital role in predicting an incident in particular situations, to monitor online for security reasons  and also for an objective driven evaluation of applications such as anomaly and intrusion detection. Presently many AI (Artificial Intelligence) based techniques are been used to detect anomalous activities among which Convolutional Neural Networks (CNN or ConvNets) using deep learning techniques has improved the precision significantly. The main aim of this project is to recommend a new model based on CNN a class of deep neural networks to detect anomaly in the video captured by surveillance cameras. This method has been trained and evaluated using the UCSD dataset and has showed an increase in the accuracy of the anomaly detection model.