1:1
1:1 Face Verification
Verify identities through one-to-one face matching for secure authentication and 2FA access control. Outstanding precision TAR 99.83% @ FMR 1E-6 on NIST report (Mar 2023).
NIST-proven accuracy, industry-leading anti-spoofing protection, and flexible deployment from edge devices to servers, all in one cross-platform SDK.
NIST FRTE 1:1
99.83%
FMR: 1 in 1 million
VISA Category, Mar. 2023
NIST FRTE 1:N
99.61%
Search in 12M Database
Mugshot Category, Jul. 2024
NIST FATE PAD
100% TRR
Ranked #1 over 82 vendors
Passive 2D Anti-spoofing, Sep. 2023
iBeta PAD
Level 2
Fully run on Mobile
100% TRR, Mar. 2022
ISO Certified
27001 & 27701
Information security & privacy information management
Face Recognition
World-class face matching accuracy, verified by NIST. From one-to-one authentication to large-scale identification across millions of faces.
1:1
Verify identities through one-to-one face matching for secure authentication and 2FA access control. Outstanding precision TAR 99.83% @ FMR 1E-6 on NIST report (Mar 2023).
1:N
Top-N search of a face database. Outstanding precision TAR 99.61% in 12M database on NIST report (July 2024).
Secure Face Templates
Converts faces into irreversible, AES-256-encrypted templates.
Face Quality Check
Checks occlusion, blur, lighting, angle, size and grayscale before enrollment or matching.
Masked Face Recognition
Recognizes faces with or without a mask; NIST validated.
Liveness & Anti-Spoofing
Presented or injected, physical or AI-generated, FaceMe detects it.
NIST FATE PAD
Ranked #1 among 82 vendors; Passive 2D Anti-spoofing, Sep. 2023
iBeta PAD
Fully run on Mobile; 100% TRR, Mar. 2022
Multi-Layered Protection
Covers presentation attacks, AI-generated deepfakes, and injection attacks with multiple layers of anti-spoofing protection.
2D Anti-Spoofing
Supports standard 2D RGB cameras with passive or active liveness for frictionless or enhanced security.
3D Anti-Spoofing
Supports 3D cameras using Structured Light, ToF, and Active Stereo.
IR+RGB Anti-Spoofing
Supports IR+RGB dual-lens cameras from multiple vendors.
Walk-Through Anti-Spoofing
Enables liveness detection without a preview screen or user interaction, ideal for IP camera-based applications.
Masked-Face Anti-Spoofing
Supports anti-spoofing and liveness detection even when users are wearing masks.
Deepfake Detection
Detects AI-generated synthetic faces and deepfake attacks.
Injection Attack Detection
Blocks attacks bypassing the camera: virtual cameras, rooted devices, VMs, API hooking.
Face & Person Attributes
Age estimation, face and person attributes, and Person Re-ID provide additional information for person analysis and search
Age Estimation
Tested on 1M images from 34 countries.
Age Verification
Determines whether a subject is 25 or older; NIST FATE benchmark (Sep 2025).
Gender & Emotion
Gender classification at 98.02% accuracy; emotion detection across 5 states (Neutral / Happy / Angry / Sad / Surprised).
Person Attributes
Detects and tracks people by body shape without a face; recognizes upper/lower clothing type, 10+ colors, and accessories (hat, bag). Enables anonymous counting, crowd monitoring, and appearance-based search.
Person Re-ID
Re-identifies the same person across cameras and time without relying on face recognition
People Tracker combines Person Re-ID, appearance-based search, and face recognition into a ready-to-deploy AI video analytics application.
Explore People Tracker →Platform & Performance
One SDK from high-end servers to low-power edge devices, accelerated on every major chipset
Broad OS Support
Windows / Windows Server, Ubuntu, Android, iOS, NVIDIA JetPack, Yocto, Debian, Ambarella Linux.
Multi-Language Support
C, C++, C#, Python, Java, Objective-C.
Broad Engine Support
OpenVINO, TensorRT/CUDA, SNPE, NeuroPilot, TensorFlow, CoreML, CVflow.
Hardware Acceleration
Offloads inference to the best available silicon on each platform: Intel CPU & iGPU (OpenVINO), NVIDIA RTX GPUs & Jetson (TensorRT), Qualcomm SNPE (DSP/GPU), MediaTek NeuroPilot (APU), NXP NPU, Ambarella CVflow, and Apple Neural Engine (CoreML).
Fast Search
Searches multi-million-person databases in milliseconds.
High-Performance Optimization
Deploy FaceMe at scale, supporting up to 45,000 people per hour on a single RTX 2000 GPU.
3-Tier Performance
Different models ranging from high-accuracy servers down to lightweight edge and mobile devices.
Compact SDK
Supports an authentication-only model variant, on-demand cloud model downloads, and three model size-reduction strategies.
Book a Demo