Case Study - machine learning platform for behavior detection system

Overview

The need for security is something increasing every time. This topic is related to many aspects like morality, economic inequality, education, and many others. As a result, the needs for a security system also increase, especially with the increasing tendecy of a homeowner left their home empty because of work, school, or other reasons. Some of the temporary external options that can fill this security task by paying security workers, housemaids or security cameras. Unfortunately, with the development of technology, there is a big shift in the availability of workers and the quality of the job. Jobs that rely heavily on physical, monotonous, or repetitive work are in demand and increase the costs needed to hire someone; so that security workers and helpers become hard choices in economic terms. With some aspect consideration of the level of trust between workers and owner who cannot be determined directly. Security cameras do not have the disadvantages that human workers pose. It's just that security camera at this time generally only limited to the presence or absence of movement, a parameter that is still too broad. So that there is no ability to get conclusions like a human could. That is why in a camera system, someone is in charge of frequently monitoring the screen and taking appropriate action. Of course, this option is not possible to be implemented on a small scale such as home in general, or even for small companies. This article contributes to improving the ability of a security system through the automation of security systems in order to recognize the security level of a place to live. The implementation discussed in this article is automation through behavior detection, face recognition, pose estimation, object detection, graphical user interface. The implementation component will be forming smart security system that can determine the security level of a place to live.

What we did