Self-Learning Analytics

Similar to human brains, the more AI-powered devices "learn" a subject, the better they analyze and respond to it. Hikvision provides methods for its deep-learning algorithms to continually improve their accuracy in any environment where deployed.

As the algorithms evolve, they gain a better understanding of objects in a scene and higher efficiency in problem-solving there.

Hikvision’s AcuSense technology, for example, is enhanced with the Learn-by-Example mode of Self-Learning Analytics. In this mode, the algorithms learn with feedback from users on alarm triggers. Users mark the false alarms generated within a particular scene and from those examples the algorithms will learn to minimize similar alarm errors in follow-up trials.

The Learn-by-Self mode enables fully automated learning by Hikvision algorithms, embedding an inference engine in the back-end device to automatically optimize algorithm accuracy in fixed cycles. This mode requires virtually no effort from users and will be available on Hikvision's video recorders in the near future.