The company is mainly aimed at developing a predictive video analytics platform “Videointellect”.

“Videointellect” is a unique and innovative product designed to solve the problem of mass theft  in retail networks and single outlets.

 The project has been under development since 2011.

Current development stage : patenting, generation and realization of smart  ideas.

The implementation of the technology is very wide: it will also be used to identify possible terrorists and criminals in places of mass gathering (such as underground systems, railway stations, checkpoints and others) as well as to increase safety in shopping and entertainment centers, sport centers and open areas.

 Methods

Our software product relies on predictive behavior technology wholly developed by our company.

The technology is based on applying statistical analysis of changes in motor behavior of people intending to commit offences.

This system of fine motor actions works automatically i.e.  without a person’s conscious effort.

Actions like these are regulated by the involuntary nervous system, which is activated by adrenaline rush and causes uncontrolled changes in motor behavior of a person intending to commit any kind of offence.

 Efficiency of the technology

The level of its efficiency is so high that it enables a detector, built on its basis, to identify the slightest change in motor behavior of a person intending to commit an offence. The program can not only detect the theft but also identify the criminal while preparing for it.

 Results

According to the results of testing the technology in the retail sector, the effectiveness of detecting exceeds 80%.

The principles of the technology

Though such a task is extremely hard to fulfil, we developed a unique approach that affords to gain great results.

A wide range of parameters that describes human’s body position, pose and movements,   undergoes analyzing and systematization with the help of multivariate statistical analysis and machine learning. As the result of this process, we get weighted coefficients of these parameters, which are used in making a final behavior model.

The algorithm for decision-making relies on the system of neural networks e i. set of neural network models. The algorithm makes a decision by averaging of their work results. Operations with neural network models can be started when balanced evaluation of complex characteristics, which have appeared as a result of analysis of motor behavior changes, is taken into account.