The People feature uses face detection and machine learning to find faces across your photo and video library, group similar faces into clusters, and let you assign names to those clusters. Once a person is named, you can browse every file in your catalog where they appear. People is currently in beta.
People is a beta feature. Results may be incomplete or imprecise, particularly
for low-resolution images, partially obscured faces, or large libraries with
significant variety in lighting conditions.
Step 1: Detect people on a source
Face detection must be run on each source before faces from that source appear in the People page. You run detection from the source in the file browser, not from the People page itself.
Open a source
Click a source in the Sources tab to open it in the file browser.
Run face detection
Right-click the source name in the sidebar, or click the options menu on
the source, and choose Detect People. You can also select individual
files in the browser, right-click, and choose Detect People to run
detection only on those files.
Wait for detection to complete
Detection runs in the background. For large sources with many images and
videos, this may take several minutes. You can continue using 1Archive while
detection runs.
Face detection works on both photos and video files. For videos, 1Archive
samples frames throughout the clip and detects faces in each sampled frame.
Step 2: Cluster detected faces
After detection, the raw face data exists but has not yet been grouped into clusters. Clustering is the process that groups similar-looking faces together so that all appearances of the same person end up in one cluster.
Open the People page
Click People in the sidebar (labeled with a BETA badge).
Click Cluster Faces
Click the Cluster Faces button at the top of the page. A status message
appears while clustering runs.
Wait for clustering to complete
Clustering analyzes all detected faces across your library and groups them
by visual similarity. When complete, a success message appears and face
clusters populate the page.
Run Cluster Faces again after detecting people on additional sources to
incorporate new faces into the existing clusters. Previously named persons are
preserved, new faces are merged into existing clusters or grouped into new
ones as appropriate.
View face clusters
After clustering, the People page displays a grid of face thumbnails. Each card represents one cluster, a group of faces that 1Archive determined likely belong to the same person. Each card shows:
- The cover photo for the cluster, the highest-quality face detection in the group.
- The person name, if the cluster has been assigned to a named person, or “Unknown” if not.
- The face count: how many individual face detections are in the cluster.
Click a cluster card to open all the files in your library where that cluster’s faces appear.
Create a named person from a cluster
Select an unassigned cluster
On the People page, click a cluster card that shows “Unknown”. The card
highlights to indicate it’s selected.
Click Create Person
Click the Create Person button in the action bar at the top of the page.
Enter a name
Type the person’s name in the dialog that appears and click Create.
Add a cluster to an existing person
If you have two or more clusters that represent the same person, you can merge them by adding one cluster to an existing person.
Select the cluster to merge
Click the cluster card you want to assign. It must not already be assigned
to a person (the name shows “Unknown”).
Click Add to Person
Click the Add to Person button in the action bar.
Choose the person
Select the target person from the list in the dialog and click Add.
Change a person’s cover photo
The cover photo is the face thumbnail that represents a person on the People page. 1Archive selects the highest-quality detection automatically, but you can change it.
Select a named person
Click the cluster card for the person whose cover you want to change.
Click Change Cover
Click the Change Cover button in the action bar. A dialog shows the top
face detections for that person.
Choose a new cover
Click the face image you want to use as the cover and click Set as
Cover.
Delete a person
Deleting a person removes the name assignment from the cluster but does not delete the underlying face detections or the files themselves. The cluster returns to “Unknown” status.
Select a named person’s cluster card, then click Delete Person in the action bar and confirm the deletion. 
