HaLO Machine Learning Algorithm is given additional features for venue and public space protection -Read More

Expansion of HaLO deep learning application for event protection


Expansion of HaLO deep learning application for event protection

July 16, 2021

HaLO (Halt and Look Once) algorithm redefines machine learning for CCTV.

Object recognition in images is now commonplace. However, the ability to turn the object classification into robotically predetermined incidents is only possible in the Merlin NEON edition. Provided that the installer set the CCTV and the examination area of the images defined to meet the requirement of ISO 62676 Initsys, Merlin can now fully automate a 'real' voice response via localised speakers and call and 'call direct' a response.

HaLO uses continuously refined training data to examine CCTV, enabling objects of interest to be detected and warnings gathered before any potential incident. Metadata referencing each image passes between modules as well as the threat manager algorithm. This data includes the count of incidents, commonality and direction.

Control of the use of HaLO can be by time/date and shortly event. So that threat evaluation is only turned on when there is an obvious danger. Such as that found when there are large groups of people present.

Merlin will highlight any threat to the Security and management team using the Apps and indications in the control room. Merlin will generate additional incidents regarding an evolving response to a threat.

This further incident may quickly clear if the threat is benign and noted as such via the responder app without calling the control room.

See Responder App.

For more information, don't hesitate to get in touch with Initsys.

Email - Admin@initsys.net for immediate attention.