New technologies are becoming more complex and sophisticated, thanks to them today we use many Internet features that greatly simplify the life of a modern person. Devices that allow you to control many processes without personal presence help people solve a variety of tasks much faster and more efficiently. Such smart devices provide comfort, coziness, cleanliness, health, provide the necessary information and, of course, organize our safety. Protecting one's home for every person is an important component of his safety, the safety of property, therefore, the market offers many options for alarm systems. Especially now, when import substitution is actively developing, Russian developers offer their innovative technologies that are used in various fields of production, including security systems.

Vadim Shaposhnikov, a student at the SFedU Institute of Computer Technology and Information Security , together with the team decided to develop innovative software using neural network technologies, which will significantly increase the level of protection of private territories.

The idea arose during the implementation of an interdisciplinary project within the framework of the subject "Introduction to Engineering", mentored by Sergey Bekezin, a laboratory assistant at the Design Office of the SFedU Institute of Computer Technology and Information Security . The project team consisted of students from the Institute of Computer Technology and Information Security and students from Manipal University (India). The project team also included a student of the Institute of Computer Technology and Information Security Victoria Plokhykh as an assistant to the author of the project and designer. 

"After analyzing the security systems market, we found out that existing alarms often turn out to be insufficiently reliable and unable to respond effectively to all possible threats. Therefore, we are planning to create a mobile application that will notify the consumer about the intrusion of intruders into the territory, accompanying the notification with an image obtained from a video surveillance camera. The application will receive images from a camera installed on the territory of the household, and when an incident occurs, it will send notifications to consumers' mobile devices so that users can see what is happening on their territory," said Vadim Shaposhnikov, hardware developer of the project.

Vadim added that in order to process the received images and analyze graphical information, images will be transferred from the camera's hardware platform to the server. Where a neural network will be launched, which is trained to analyze graphical information and determine the presence of a person. The neural network algorithms will be optimized to ensure high accuracy and processing speed.

Vadim Shaposhnikov noted that, unlike standard alarms, the system he is developing has a number of advantages, including:

- a more intelligent and adaptive threat response algorithm;

- advanced functionality that allows you to detect various emergency situations, which includes the ability to flexibly adjust the sensitivity and operating modes of sensors to suit the characteristics of the premises, integration with high-performance IP cameras for round-the-clock monitoring, the ability to connect sirens, light indicators and other sirens, remote access to video streams and the ability to view the archive through a mobile application;

- Remote monitoring and management via the mobile app;

- self-learning and improving the efficiency of the neural network.

"Our product is suitable for private households, apartment owners and tenants, as well as owners of small businesses. As a result, customers receive a modern, flexible and high—tech security system capable of providing comprehensive protection of their territory," the author of the project said.

A beta version of the system has already been developed. The project team plans to improve its hardware platform and increase the level of data transfer security, in addition, the team plans to create a product variant for sale on marketplaces. According to Vadim Shaposhnikov, the grant won will be aimed at developing new neural network algorithms, purchasing more efficient hardware and testing it, advertising and implementing the system on real objects.