Skip to main content

Overview

The Enrollment SDK builds an Applicant — an enrolled person's biometric representation (images, templates, and quality) — from face, iris, and fingerprint (print) modalities, and matches against it.

Capabilities​

  • Image processing — decode and prepare face, iris, and fingerprint images for the pipeline.
  • Template extraction — produce a per-modality template for an Applicant.
  • 1:1 verification — compare two applicants and return a similarity score.
  • 1:N identification — search a probe against a gallery (ModalitiesGallery, IndexedModalitiesGallery).
  • Quality — per-modality image quality scores (fingerprint, iris, face) to decide whether an image is suitable for further processing.
  • Fingerprint capture — slap capture with algorithmic segmentation into individual fingerprints, per-finger position, and duplicate/flattening checks.
  • Server execution — run image, print, face, and iris operations on a provided server process (enrollment_server) instead of in-process.
  • Extensibility — several features can be adapted to your needs by implementing the provided interfaces (see the SDK's "Extending SDK" reference).

Modalities​

ModalityPipeline
Facedetection, crop, validation, quality (OFIQ), template extraction
Irisdetection, segmentation, position, template extraction
Fingerprint (print)slap capture, algorithmic segmentation, per-finger position, template extraction

Palm is not covered by the Enrollment SDK.

Supported platforms​

  • Linux — x86_64
  • Windows — x86_64 or x86 (on x86, iris/face/print embedding are not supported)
  • Android — 5.0 or higher (API level 21)

For the exact runtime prerequisites (glibc, AVX2, libusb), see Runtime requirements.

Connectors​

A C++ core with Java and C# connectors; runnable samples ship for all three. See the API Reference.