The solution consists of a video camera, recording the building’s entrance; the Enterprise Node, which could be deployed on any laptop with internet access; a containerized smart contract with a neural network responsible for facial recognition; and a frontend application providing functionality for viewing attendance statistics, adding new employees, and interacting with the smart contract and Node.
A live video stream from the camera is fed to the Frontend Application for the security guards to view directly, if necessary.
The video stream is broken down into images, which are sent to the Node as a series of transactions.
The Node passes the transactions to the smart contract.
Whenever the smart contract recognizes a face, it records on the blockchain that the relevant employee has entered, with their personal ID and a timestamp. The application also records images and timestamps of unknown faces, in order to keep track of strangers or in case the neural network fails to recognize an employee.
The Node stores these transactions, aggregates daily and monthly statistics, and broadcasts them to the network. The Frontend Application displays this information and statistics.
Since all nodes in the network share the same transaction history, altering a record would require that a majority of nodes agree a fraudulent consensus and rewrite multiple blocks — ensuring protection against attacks. Similarly, as transactions are appended to the existing history only if consensus is agreed beforehand by a majority of network nodes, fraudulent transactions on the private blockchain are virtually impossible.
The use of blockchain leads to reduced paperwork and dramatically decreased human error and administrative costs.