"Beeline" has developed a specialized neural network that assists in the search for missing persons: the platform is called "Beeline AI - People Search."
The solution is designed to simplify the work of the search-and-rescue team "". Since 2018, this team has been using drones for search operations conducted in forests and urban industrial areas. However, analyzing images obtained from drone cameras requires a large number of volunteers, and it takes a lot of time.
The neural network "Beeline AI - People Search" is intended to automate the photo processing workflow. It is claimed that specialized algorithms can reduce the time spent on reviewing and sorting the received images by two and a half times.
The platform employs convolutional neural network technology, which enhances the efficiency of machine vision tools. The neural network was trained on actual image sets, and tests have shown that the model's accuracy on test images approaches 98%.
The primary task of "Beeline AI - People Search" is to sort out "empty" and uninformative photos that definitely do not contain a person or attributes indicating that a person was present at that location. This allows the analysis team to immediately focus on potentially fruitful images.
The system can adapt to various conditions. It accurately detects objects from heights of 30–40 meters as well as from an altitude of 100 meters. The neural network is also capable of processing images with a high level of visual "noise"—trees, natural landscapes, twilight, etc.
"Potentially, the neural network can locate people and objects in all search areas, such as woods, swamps, fields, and cities, regardless of the season or the clothing of the person, as the algorithm is configured to work year-round and can potentially recognize non-standard body positions in space, for instance, a person sitting, lying down, or partially concealed by foliage," says "Beeline."
Source: 3dnews.ru
