Skip to main content

Embedded Biometrics

API & technical documentation

API documentation and integration samples for this solution are distributed through the Innovatrics Customer Portal. Contact sales@innovatrics.com to request access.

Innovatrics Embedded Biometrics delivers Innovatrics' biometric algorithms, SDKs, and components for on-device deployment. Rather than a complete end-user system, it provides reusable building blocks — face, fingerprint, iris, and palm — that OEM partners and system integrators embed directly into their own hardware and applications. Components run locally on the device — handhelds, kiosks, PoS terminals, cameras, AI boxes, and access-control terminals — without dependence on a server or cloud back-end, and can equally support edge-to-server architectures.

The SDKs are built on Innovatrics' NIST top-ranked algorithm technology and shipped as embedded-optimized builds, tuned for ARM/NPU targets with a small footprint, on-device liveness detection, and high accuracy on the edge — no cloud dependency and no vendor lock-in.

Documentation​

This solution's documentation is organized as follows:

  • Solution Composition — the component architecture, the on-device SDKs, and the Biometric Identification Service, with the full SDK comparison
  • Use cases — end products you can build (on-device identification, server-side identification, …) and how to assemble them
  • Features — individual biometric capabilities (identification, verification, enrollment, liveness, …), independent of any single SDK
  • Components / SDKs — the SDKs that implement the features: SFE Toolkit (face, iris, palm), Enrollment SDK (fingerprint, face, iris), IDKit and ANSI/ISO SDK (fingerprint), plus the Biometric Identification Service (server)
  • Standards & Certifications — the NIST and ISO benchmarks the algorithms are evaluated against

What it does​

The SDK suite provides modular biometric capabilities that can be composed to match specific product requirements:

  • Facial recognition — matches faces from cameras or other input devices against local databases or watchlists for real-time identification on edge devices; algorithms consistently top-ranked in NIST FRTE 1:1 and NIST FRTE 1:N evaluations (SFE Toolkit, Enrollment SDK)
  • Fingerprint recognition — high-speed, high-accuracy fingerprint detection and matching with ANSI/ISO interoperability; algorithms top-ranked in NIST PFT III and NIST MINEX III (Enrollment SDK, IDKit SDK, ANSI & ISO SDK)
  • Iris recognition — contactless, high-precision identification using deep learning; iris patterns remain stable for life with no re-enrollment required, yielding one of the lowest false match rates of any biometric modality; top-ranked in NIST IREX 10 (SFE Toolkit, Enrollment SDK)
  • Palm recognition — contactless palm capture from any standard camera, unaffected by temperature, health conditions, or seasonal changes; supports both 1:1 verification and 1:N identification (SFE Toolkit)
  • Presentation Attack Detection (PAD) — passive liveness detection for face and palm that confirms physical presence and blocks printed photos, screen replays, and 2D/3D replicas; ISO 30107-3 Level 1 & 2 certified, 100% spoof detection in Level 2 testing, results in under one second
  • Age & gender estimation — automated age and gender estimation from facial images, supporting age-restriction policies and minors detection enforced at the device level (SFE Toolkit, Enrollment SDK)
  • Auto-capture (face and palm) — guided capture components for Android and iOS that trigger acquisition once framing, distance, and quality thresholds are met, reducing operator effort and improving first-attempt capture quality (SFE Toolkit)
  • Camera stream processing — direct analysis of one or multiple concurrent camera feeds for real-time video analytics or access control on embedded hardware (SFE Stream Processor)

Who it's for​

Embedded Biometrics serves two integrator audiences:

  • Hardware vendors (OEMs) — manufacturers embedding biometrics into devices produced and sold in volume. One modular stack, low footprint, and per-device licensing.
  • System integrators — teams delivering enterprise or government projects that need edge-side biometrics within a larger system, with clean SDK interfaces and no vendor lock-in.

Target use cases​

The SDKs are used across a wide range of hardware and application contexts:

  • Access control and time & attendance — enterprise entry terminals, time-clock readers, facility management systems, stadium and event security gates
  • Handheld and mobile — biometric scanners, industrial rugged devices, smartphones and tablets
  • PoS and fintech — payment authentication at point-of-sale, biometric payment cards, ATM access, fraud prevention at transaction terminals
  • Smart home and IoT — smart locks, home security cameras, appliances requiring user authentication with a focus on local processing and privacy
  • Automotive — driver authentication, keyless entry, personalized vehicle settings, in-cabin occupant monitoring

Deployment options​

  • Fully on-device / offline — recognition, liveness, and matching run entirely on the device with no cloud dependency, suited to air-gapped and privacy-sensitive environments
  • Hybrid (edge + central 1:N) — on-device capture and liveness, with centralized identification across a fleet via the server-side Biometric Identification Service
  • On-device SDKs run on embedded Linux, Android, and PC (Windows / Linux); server-side components run on Linux
  • Hardware-agnostic — compatible with a wide range of cameras, fingerprint sensors, iris scanners, and palm capture hardware from major vendors
  • Modular by design — individual modalities and capabilities can be deployed independently without pulling in unused functionality
  • Flexible licensing tailored to device-level, project, and fleet deployments, with a free evaluation and a fixed-price integration package to get started — contact sales@innovatrics.com for options

Integration approach​

The SDKs expose biometric functions through documented APIs and ship with integration samples covering common use cases. Integration is designed for minimal friction in both device-level and server-side contexts.

Key integration touchpoints:

  • Native SDKs and language bindings — C/C++ core libraries for direct embedding into device firmware and native applications, with platform-specific bindings including Java/Kotlin (Android), .NET, and Python
  • Documentation and samples — case-specific integration samples and API documentation provided through the Innovatrics Customer Portal; contact sales@innovatrics.com to request access
  • Hardware-agnostic capture — no sensor vendor lock-in; the SDK is validated against a broad set of commercially available capture hardware

Integration support​

Innovatrics provides direct engineering support throughout the integration lifecycle:

  • SDK documentation and code samples covering common device integration patterns
  • Direct access to Innovatrics engineers for integration questions and performance tuning
  • Performance validation and chipset-specific optimization on the partner's target hardware during evaluation and integration
  • Long-term co-development partnership — engineering alongside the partner across an Evaluate → Integrate → Scale model
  • 20+ years of algorithm development experience across government, enterprise, and consumer biometric deployments

Innovatrics Academy​

Innovatrics Academy offers online training courses covering integration, configuration, and operation of Innovatrics solutions. Some courses are publicly available at no cost; others are available upon request through your Innovatrics sales contact, including courses not listed in the public catalog.

License Agreement​

Use of this solution is subject to the Innovatrics License Agreement.