Skip to main content

Auto-capture

Auto-capture automatically captures a well-framed, quality-checked face or palm from a live camera stream, so the image handed to enrollment or recognition is good on the first try. It is an optional add-on, provided by the DOT SDK on Android, iOS, and Web.

How it works

Auto-capture wraps the front of the pipeline (capture, with on-the-fly detection and quality checks): it watches the camera stream, guides the user, and emits a frame only when framing, distance, sharpness, and lighting pass. From there the normal flow continues — see Identification (1:N), Verification (1:1), and Enrollment. It can be used for all of them.

Capture quality

The component selects a frame that meets a configurable quality bar, trading user effort against image quality through quality providers (face-comparison, passive-liveness, ICAO), and guides the user through a "stay still" phase. The detailed quality providers, capture process, and UI recommendations are documented once for the DOT Face Auto Capture component — see Face Auto Capture in the IDV Toolkit documentation.

Availability

SDKDOT (separate from SFE Toolkit / IDKit)
PlatformsAndroid, iOS, Web
ModalitiesFace, palm
StatusOptional, billable add-on
DocumentationFace Auto Capture overview · face per platform: Android, iOS, Web; palm: Android, iOS

Key considerations

  • Mobile / web only — on embedded Linux terminals there is no auto-capture component; the integrator supplies camera frames directly. An integrator can implement a similar auto-capture loop with SFE Toolkit primitives (detection, quality and pose attributes, face area) to gate which frames enter the pipeline.
  • Quality at the source — better capture improves downstream liveness and matching accuracy.