is a system for face verification. It has built in appearance/age change factor to account for hair styles, facial hair, eye wear as well as the passage of time. Facelab is based on deep neural networks (deep learning) which makes it a solution that uses the latest advancements in artificial intelligence.

Facelab uses convolutional neural network to extract traits from every face that is analyzed. The extracted features are used to compare two faces. To achieve this, another neural network is created, which is a siamese neural network. This network compares the features of two faces and determines whether they are one and the same person.

Thanks to the extraction of features, Facelab is able to match the image of a person’s face to a previously defined database of photos of people and thus is able to identify the person in the picture.

The system has been designed in such a way that it is insensitive to interference such as changes in lighting, scale, sharpness, color or angle at which the face is observed. That is why it is highly effective in face verification.

Sample implementation scenarios

Confirmation of the client’s identity

In the customer service process, it is possible to check whether the person who passes the identity verification process is the owner of the document.   The following example compares facial features from a photo from a document (on the left) and a photo taken with a webcam. 

Fraud detection 

The system compares the client’s image from each process with all images available within the instance and examines whether the identity has been verified using a document containing different data but the same or similar image. In this way the system can detect fraud attempts using mass-produced false documents with a photo of the same person.

Identification of people entering a protected area

The tool can be used for automatic verification of visitors and provide support for the reception staff. A person entering a building can scan his ID document using a self-verification stand. He is guided step by step through the stages of the verification. The system retrieves the image of the person from the camera and compares it with the picture on the document. There is also possibility verify the credibility of the document (integration with idenTT Verification System required). By the time the visitor approaches the service desk, his data will already be in the system

Our system can:

  • Save information about visitors
  • Send the data of the person entering to the receptionist
  • Notify building security of unauthorized entry
  • Verify whether the person who use authentication card is the actual owner of the card
    (integration with proximity reader)

Liveness detection

The solution can also determine whether the real face of a person is presented in front of the camera. At present, the solution is based on asking the client to perform certain gestures in a certain sequence (head movements in the indicated direction and blink) and verify the correctness of the task.

Please visit facelab.io, to see the demo version of the solution.

If you have any questions, please feel free to contact us any time.

11 + 3 =