face-enroll

To enroll a face using a photo, the self-enroll flow submits an image to the /face/enroll HTTP endpoint. The system validates the submission, applies lowercase labels, and adds the face to the lamp's recognition roster.

Yes, face enrollment supports a camera capture flow that uses the /camera/snapshot HTTP endpoint. This self-enroll path takes a live snapshot, validates that a face is detected, and adds the captured face to the lamp roster.

If no face is detected during camera enrollment, the system returns an error and the enrollment fails. The skill handles this edge case along with camera offline errors to prevent invalid entries in the face recognition roster.

You can manage the face recognition roster using maintenance operations that check status, remove specific enrolled faces, or reset the roster. These maintenance tasks use HTTP endpoints like /face/status to update the enrollment set.

Yes, the face enrollment system includes a familiar-stranger prompt flow. This path manages memories for friends and strangers by enrolling faces with appropriate labels and Telegram metadata where applicable.

Lamp face enrollment and management use HTTP endpoints including /face/enroll for adding faces, /camera/snapshot for camera capture, and /face/status for checking roster status. These endpoints support validation and lowercase labeling.