A team of scientists from Russia and Uzbekistan has created a neural network that can automatically recognize street violence in video surveillance footage, according to TASS.
According to an associate professor at the South Ural State University, external surveillance cameras generate a large stream of data, which overloads existing neural networks.
The scientists used fuzzy logic in creating the neuron together. The network allowed them to successfully solve the problem of this excess load on the cameras. As a result, the developed neural network was able to record more than 98% of violence instances captured by street surveillance cameras.
Additionally, the university mentioned that the new neural network is intended to increase the level of security in populated areas. It can be successfully integrated into the Smart City system.
