Skip to content

Face Recognition & Matching

The AI Photo Gallery’s core feature is face recognition — the system detects every face in your event photos, groups them by person, and lets attendees find all their photos with a single selfie.

  1. Detection — AI scans each uploaded photo and identifies individual faces.
  2. Grouping — Detected faces are compared and clustered. Faces that look like the same person are placed in the same group.
  3. Matching — when an attendee takes a selfie, the system compares it against all face groups.
  4. Results — the attendee sees every photo where their face appears.

Go to AI Gallery → Dashboard to see face recognition metrics:

MetricWhat It Shows
Total GroupsNumber of unique people detected across all photos
Matched GroupsGroups where the person has been identified (selfie matched)
Unmatched GroupsGroups where the person hasn’t been identified yet
Total FacesTotal individual face detections across all photos
Average Photos Per PersonHow many photos the average attendee appears in
Match RatePercentage of groups that have been successfully matched

Go to AI Gallery → Face Groups (or the face management section):

Each group shows:

  • Thumbnail — a representative face from the group
  • Photo count — how many photos this person appears in
  • Status — matched (identified) or unmatched
  • Attendee name — if matched, shows the attendee’s registration name

Sometimes the AI creates two groups for the same person (e.g., different lighting or angles). To merge:

  1. Select the duplicate groups.
  2. Click Merge.
  3. All photos from both groups are combined under one identity.

If the AI grouped two different people together:

  1. Open the group.
  2. Select the incorrectly grouped photos.
  3. Click Split to create a new group from the selected photos.

Attendees find their photos using a selfie:

  1. Attendee opens the AI Photo Gallery link (from email, WhatsApp, or event website).
  2. They take a selfie or upload a photo of themselves.
  3. The AI matches their face against all detected groups.
  4. They instantly see every photo they appear in.
  5. They can download, share, or save their photos.

For events that collect attendee photos during registration:

  • The system can automatically match registration photos against face groups.
  • Attendees don’t even need to take a selfie — their photos are ready when they visit the gallery.
  • This works best when your registration form includes a photo upload field.

For each photo, you can see which attendees appear in it:

  1. Open a photo.
  2. View the Faces Detected panel — shows all faces found in this photo.
  3. Each face links to its group (and to the matched attendee, if identified).

This is useful for:

  • Verifying that key speakers and VIPs were properly detected
  • Checking that group photos correctly identify all participants
  • Quality-checking the AI’s face detection accuracy

Face recognition accuracy depends on photo quality:

FactorImpact
LightingWell-lit photos have much higher detection rates
Face angleFront-facing faces match best; extreme profiles may be missed
ResolutionHigher resolution = better face detection
OcclusionMasks, sunglasses, or hands covering faces reduce accuracy
DistanceFaces that are too small in the frame (far from camera) may not be detected

  • Use professional photographers — well-lit, high-resolution photos dramatically improve face detection accuracy.
  • Capture front-facing photos — instruct photographers to capture people facing the camera when possible, not just candid side angles.
  • Review face groups — spend 10–15 minutes after processing to merge obvious duplicate groups and split any incorrect groupings.
  • Promote selfie retrieval — the gallery is only valuable if attendees use it. Send a WhatsApp campaign with the gallery link right after the event.
  • Check VIP coverage — verify that your keynote speakers, sponsors, and VIP guests are correctly detected and grouped. Manually merge groups if needed.