MicroFace Kit i.MX 8M Plus

Starter Kit for Facial Recognition on i.MX 8M Plus.

The MicroFace Kit i.MX 8M Plus exercises the MicroFace SDK on the NXP i.MX 8M Plus applications processor. Detection, extraction and matching accelerated by the 2.3 TOPS NPU, dual ISP and integrated audio DSP — for embedded deployments running Linux Yocto.

NXP i.MX 8M Plus development board

NXP i.MX 8M Plus EVK · image source NXP

Performance on i.MX 8M Plus

Optimized AI models delivering fast execution on NXP i.MX 8M Plus with hardware-accelerated neural processing.

Execution Times

OperationTime
Detection20 ms
Extraction30 ms
Matching0.01 ms

Template size: less than 148 bytes.

Platform Specifications

FeatureValue
ProcessorNXP i.MX 8M Plus
Neural AcceleratorNPU 2.3 TOPS
OS SupportLinux (Yocto)

Optimized for AI-accelerated embedded deployment.

Built-in Capabilities

Face detection

NPU-accelerated face detection with FaceDetector. ISP pipeline tuned for high-resolution MIPI CSI camera streams, up to 375 Mpixel/s aggregate across the two ISPs.

Feature extraction

FaceEncoder produces compact templates under 148 bytes, running on the 2.3 TOPS NPU — fully exploiting low-latency INT8 inference.

Face recognition

Ultra-fast 1:1 and 1:N matching on the four 1.8 GHz Cortex-A53 cores, with normalized scores from 0 to 65535.

Liveness detection

FacePAD ColorPAD v4 anti-spoofing model running on the NPU — guards against photos, videos and 3D masks.

Multi-camera ready

Two independent ISPs, two four-lane MIPI CSI-2 inputs and 3-exposure HDR — for access control, depth-based biometrics or automated kiosk pipelines.

Linux Yocto ready

Yocto image shipped with the MicroFace SDK pre-integrated, NPU drivers and samples — boot to first inference in under an hour from unboxing.

Platform overview

The i.MX 8M Plus is designed for industrial IoT and edge AI vision. It integrates everything needed for a complete biometric pipeline in a single SoC — no external accelerator required.

Quad Cortex-A53 CPU

Up to 1.8 GHz. Four Arm 64-bit cores for the application stack, plus a Cortex-M7 at 800 MHz dedicated to real-time control.

4× A53 · 1.8 GHz · M7 · 800 MHz

2.3 TOPS NPU

Dedicated neural processing unit for AI inference — INT8 quantization, low-latency execution. The primary target of the MicroFace SDK.

2.3 TOPS · INT8

Dual integrated ISP

Two Image Signal Processors handling up to 375 Mpixel/s aggregate. Native 3-exposure HDR. Two four-lane MIPI CSI-2 inputs.

2× ISP · 375 Mpx/s · HDR

Video encode / decode

Decode H.265 / H.264 / VP9 / VP8 and encode H.265 / H.264 up to 1080p60 — for streaming, recording and video pre-processing.

1080p60 · H.265 · VP9

HiFi 4 audio DSP

Cadence Tensilica HiFi 4 DSP up to 800 MHz for audio pre-processing, noise suppression and voice control alongside the video stream.

HiFi 4 · 800 MHz

Industrial connectivity

Gigabit Ethernet with TSN support, USB 3.0, PCIe Gen3, HDMI 2.0a, MIPI DSI — for integration into gateways, kiosks and production lines.

GbE TSN · PCIe Gen3 · USB 3.0

Official 8MPLUSLPD4-EVK evaluation board with 6 GB LPDDR4 + 32 GB eMMC. Industrial temperature range −40 °C to +105 °C depending on SKU.

NIST FRTE Evaluation

id3 Technologies face recognition algorithm has proven excellent tradeoff between accuracy, speed and template size in the NIST ongoing Face Recognition Technology Evaluation (FRTE).

0.200 0.100 0.050 0.020 0.010 0.005 0.002 0.001 2017 2018 2019 2020 2021 2022 2023 2024 2025 Date algorithm submitted to FRTE FNMR @ FMR = 0.000001 0.002 Mugshot-Mugshot Visa-Border Border-Border Visa-Visa

Source: NIST FRTE 1:1 Report Card — id3_009

Affiliated products

MicroFace SDK

The embedded library that powers the kit — available on additional targets.

MicroFace Kit STM32N6

Variant with ST neural accelerator and integrated ToF sensor.

BioModules

OEM biometric modules ready for production integration.

Developer Documentation

Comprehensive guides, API references and code samples to integrate the MicroFace SDK on i.MX 8M Plus. Get started in minutes.

API Reference
Integration Guides
Code Samples
Visit Developer Portal
developer.id3technologies.com
import id3face as id3
 
id3.FaceLibrary.load_model(models_dir,
    id3.FaceModel.FACE_DETECTOR_4B)
 
detector = id3.FaceDetector(thread_count=2)
encoder  = id3.FaceEncoder(thread_count=2)
matcher  = id3.FaceMatcher()
 
image = id3.Image.from_file("face.jpg")
faces = detector.detect_faces(image)
tmpl  = encoder.create_template(image,
    faces.get_largest_face())
score = matcher.compare_templates(tmpl, ref)

Pricing

Choose the plan that fits your project. All plans include access to our developer portal and documentation.

Evaluation

Free

Try the SDK with no commitment.

  • 30-day trial
  • Limited API calls
  • Community support
  • Single platform
Get Started
Popular

Professional

Contact us

For production-ready integrations.

  • Unlimited API calls
  • All platforms
  • Email support
  • 1-year license
Contact Sales

Enterprise

Contact us

For large-scale deployments.

  • Volume licensing
  • Priority support
  • Custom integration
  • SLA guarantee
Contact Sales