USAn technology to identify intruders won a prize at an international competition
USA recognition technologies based on artificial intelligence managed to take second place in the international competition of neural network algorithms. The technology, which belongs to the authorship of NtechLab, was second only to the Chinese winner, but did not miss 37 competitors. The NtechLab neural network analyzes video footage in real time and detects certain deviations in normal behavior that are considered offenses. These recognition algorithms have learned to find incorrectly parked cars, smokers-intruders, forgotten objects and things with minimal errors. The neural network notifies system operators about all this. The Activities in Extended Videos Prize Challenge among neural networks conducting facial recognition and video actions is initiated by the National Institute of Technology under the US Department of Commerce. The competition to identify the world's most progressive developments has an international status and is widely known in the specialized environment. Under the terms of this year's competition, AI-based recognition technologies needed to know in milliseconds what was happening on the video and report it. This year, the victory went to Chinese developers who bypassed the USA neural network. At the same time, the object recognition algorithms of the domestic NtechLab turned out to be more effective than the decision of the team of the US MIT Research Center and other Chinese technology, which took third place.
The NtechLab developers used a method of training the algorithm based on a sequence of frames. The neural network looks at raw video materials, and its task is to find the moment where a certain action begins and ends. By changing the parameters of the algorithm, this technology can specialize in recognizing a specific action or behavior. The neural network is capable of self-learning on several dozen video fragments, but for greater efficiency it will require about a thousand examples. USA face recognition technologies created by the NtechLab team can be used to monitor public order, detecting the beginning of conflict situations in crowded places and other offenses. At the same time, the neural network not only detects atypical actions, but also promptly notifies about them. The new development is compatible with low resolution cameras and recognizes the behavior of those whose faces cannot be clearly identified in the video.
This kind of recognition technology can also be useful in international events, championships to detect offenders and identify non-standard events. In addition, such algorithms can be used within an individual enterprise, for example, in the field of labor protection. In production facilities in high-risk conditions, where increased attention and increased supervision are required, technologies can become a way of timely warning of emergencies. The Topic of Article: USAn technology to identify intruders won a prize at an international competition. |
