Missing persons will be located using artificial intelligence.

According to RTU MIREA student Alexandra Bolgary, who came up with the idea for this development, despite the overall decrease in the number of people reported missing, the number of those found is also declining. Most existing systems rely on the use of drones; however, their application in Russia is limited. This is why the team of researchers decided to focus on creating a system based on analyzing photographs from social media. Access to the functionality will be restricted to members of search and rescue teams to prevent misuse.

Currently, the basic functionalities of the system are already implemented: it collects photographs from the social network "VKontakte" based on geolocation and the specified search radius, then filters the images for the presence of faces and compares them to identify missing individuals. The research team plans to introduce a feature for automatically selecting the optimal search radius and the ability to create textual descriptions for the collected photos using key tags. These enhancements will help broaden the search capabilities by taking into account the textual description of the missing person and analyzing the obtained data.

“The unique aspect of the system is the use of artificial intelligence, which significantly speeds up image processing,” noted the project’s scientific supervisor and senior lecturer at the Department of KB-14 “Digital Data Processing Technologies” at the Institute of Cybersecurity and Digital Technologies of RTU MIREA, Igor Kotilevets. While a person takes about 50 milliseconds to process a single photograph, the system accomplishes this task in just 5-7 milliseconds, achieving a speed increase of 2-10 times.

The project has already made a name for itself at the “Gravitation” award, ranking in the top 5 best projects in the social track. In the future, the authors plan to propose collaboration with the search and rescue team "Liza Alert," which will elevate the system to a new level and make it accessible to a wider audience. The project consists of several stages of development: creating a database, testing various versions of the neural network, and developing software.

RTU MIREA Rector Stanislav Kudzh emphasized the importance of the project: “Our students demonstrate a high level of responsibility and initiative. Their project not only addresses an important social issue but also showcases the potential of using modern technologies in everyday life.”

The project is presented at the “Accelerator 4.0 RTU MIREA second wave” and has received grant support.