Money

Neural Network-based analyzer of the correctness of financial transactions by employees and the adequacy of visitor behavior. 

Behaviors that can be monitored:

• Stealing money from the table

• Leaving bank cards unattended

• Leaving a bank safe deposit box or safe open

• Payment in cash

• Payment via ATM

• Counting visitors

• Waiting time

• Service time

• Queue length

• Weapon detection

• Visual alarm

• Quick review of customer settlement events

• …

Examples of algorithms:

Stealing money: The neural network identifies all “money” objects in a defined area (e.g., table) and waits for the appearance of the “money” class in other defined areas (vacuum cleaner, safe, etc.). If, after the disappearance of the “money” object in the first zone for N seconds, the “money” object does not appear in any other defined zone, the algorithm triggers. This eliminates the disappearance of banknotes.

Leaving bank cards unattended: The neural network identifies the “card” object and the “person” object in a defined image area from the video camera covering the workstation / negotiation table. If the “card” object remains in the field of view of the camera(s) and, at the same time, the “person” object is absent in the defined area for a unified period of time, the algorithm triggers. The permissible time for the absence of the “person” object in the frame is N seconds (configurable parameter).

Leaving a bank safe deposit box or safe open: The neural network identifies the “open cabinet” class based on the position of the open door. After that, it determines if there is the “person” class in the defined image area covering the employee’s workstation at the cabinet, if the class is absent for more than N seconds, the algorithm triggers.

Can work both in large banks and mini-offices, as it operates on NEURO-CAPSULE, which does not require space, cooling, or maintenance.

More detailed…