Self Learning

The Self Learning tab shows how the AI resolves unidentified camera detections into real member profiles. Over time, the system accumulates visit evidence, compares movement patterns against check-in data, and automatically links anonymous detections to known members. Staff can review borderline cases and manually approve or reject matches.


Decision flow

The How profiles resolve diagram shows the pipeline each unidentified profile moves through:

StageOutcome
Successfully linked to member profileAutomatically Linked
Has visit/swipe logs but not enough info to link (e.g. missing name)Unable to Link
Enough visits + validation checks failedPending Review
No single member correlationNo Match (Monitoring)

The Decision Flow diagram also shows the count of profiles at each stage, the last time the pipeline ran, and a summary description: "Automatically links unidentified people to real members by correlating visit patterns with check-in data. Useful when members don't have valid profile photos or their photos are of low quality."


Each status card displays a count, a percentage of total profiles, and a short description:

StatusDescription
In PipelineProfiles still gathering evidence from repeat visits and check-ins. Shows an Average progress bar indicating how close profiles are to the required visit threshold (e.g. 3/8).
Automatically LinkedSuccessfully matched and linked to a real member profile by the AI.
Pending ReviewStrong timing match found, but one or more validation checks did not pass. Requires staff review.
Unable to LinkEnough visit data exists but the system lacks sufficient information (e.g. missing member name) to complete the link.
RejectedMismatches that staff have explicitly reviewed and ruled out.
MonitoringProfiles being actively monitored — typically guests, staff, or individuals with ambiguous visit patterns.

Profile table

Below the summary cards, tabs filter the table by status (e.g. In Pipeline 635, Pending Review 87, Automatically Linked 73, Unable to Link 5, Rejected, Manually Approved 1, Monitoring 646).

A Duplicates toggle and search field are available to filter the table by name, ID, or person ID.

Table columns

ColumnDescription
FacesDetected face thumbnails associated with this profile (may show multiple captures)
Person / MemberThe matched or candidate member profile, with name, status, and member ID
VisitsNumber of times this individual has been detected
Avg LagAverage time difference (in seconds) between camera detection and check-in swipe — a negative value means the camera detected them before they swiped
ScoreAI confidence score for the match
StatusVisit threshold progress (e.g. 3/8 visits required)
UpdatedWhen this profile was last updated

Related pages