Fortinet black logo

Administration Guide

Identifying faces

Identifying faces

A face cluster is a collection of images of the same person detected by face recognition AI. Face clusters are categorized as either known or unrecognized (new) faces.

Reviewing new faces

All unrecognized face clusters are categorized as new faces. If you know the person, you can link the face cluster to that person's record in the face recognition database so that FortiRecorder can recognize the face.

To link a new face cluster to a person

  1. Go to Face Recognition > Face Cluster > New Faces.
  2. Select the time range from the dropdown menu to display face clusters captured in that time period.

  3. Select a face cluster to open all images that the AI has grouped as the same person.
  4. Confirm that all images contain the same person. If an image does not show the same person, then hover over the image and click the trash can icon in order to correct the AI.

    Caution

    Verify every face in the face cluster. If faces are not correctly identified, and you do not correct it, then AI identification will be less accurate.

  5. Either:

    • From the Search User dropdown menu, select an existing person to whom the face cluster belongs. The five closest matches appear.
    • Click New User and enter their person's name, role, and department and then click Save. For details, see Configuring a department and role.
  6. Click Link all to user.

    The individual's face cluster is now a known face. All future face images are analyzed by the AI and categorized by their face cluster.

Reviewing known faces

Known face clusters contain images of people that are recognized by face recognition AI, or that you manually linked to an entry in the face recognition database.

To edit known face clusters

  1. Go to Face Recognition > Face Cluster > Known Faces.
  2. Select the face cluster for a person to open that record. For details, see Configuring a department and role.
  3. Hover your cursor over the profile picture and click More.

    All the face clusters manually linked to the user are displayed.

Identifying faces

A face cluster is a collection of images of the same person detected by face recognition AI. Face clusters are categorized as either known or unrecognized (new) faces.

Reviewing new faces

All unrecognized face clusters are categorized as new faces. If you know the person, you can link the face cluster to that person's record in the face recognition database so that FortiRecorder can recognize the face.

To link a new face cluster to a person

  1. Go to Face Recognition > Face Cluster > New Faces.
  2. Select the time range from the dropdown menu to display face clusters captured in that time period.

  3. Select a face cluster to open all images that the AI has grouped as the same person.
  4. Confirm that all images contain the same person. If an image does not show the same person, then hover over the image and click the trash can icon in order to correct the AI.

    Caution

    Verify every face in the face cluster. If faces are not correctly identified, and you do not correct it, then AI identification will be less accurate.

  5. Either:

    • From the Search User dropdown menu, select an existing person to whom the face cluster belongs. The five closest matches appear.
    • Click New User and enter their person's name, role, and department and then click Save. For details, see Configuring a department and role.
  6. Click Link all to user.

    The individual's face cluster is now a known face. All future face images are analyzed by the AI and categorized by their face cluster.

Reviewing known faces

Known face clusters contain images of people that are recognized by face recognition AI, or that you manually linked to an entry in the face recognition database.

To edit known face clusters

  1. Go to Face Recognition > Face Cluster > Known Faces.
  2. Select the face cluster for a person to open that record. For details, see Configuring a department and role.
  3. Hover your cursor over the profile picture and click More.

    All the face clusters manually linked to the user are displayed.