Smart Surveillance Tool

Abstract

Most existing video surveillance systems require monitoring of the footage 24/7 in order to discover
abnormalities and crisis situations; this is labour intensive and inefficient. Through this paper, a tool that uses
various viable technologies and methods that can augment the existing surveillance system has been discussed.
The aim is to develop a surveillance tool that focuses on detecting, recognizing and indexing humans.
Additional features to detect anomalies using object and text detection, have also been implemented. The tool is
designed to be easily scalable and to work on live camera feed