The AMT-FAPVR-21 is a multimodal, self-contained device for recognizing faces and palms. It has internal algorithms for face and palm recognition so that it can complete the cycle of biometric identification. This involves scanning faces and palms, capturing Near-Infrared (NIR) grayscale and true-colour RGB images, analyzing attributes of faces and palms, extracting templates, and matching biometric data. These deep learning-based built-in algorithms in the module work at high speeds for enrolling and authenticating faces and palms, thus enabling integrators to take advantage of the computing power capabilities offered by this module. This may result in lower dependence on computing resources available locally while also making compatibility with different hardware platforms possible. Based on AI computer vision technologies, the built-in face and palm algorithms provide highly secure liveness detection, preventing forgery hacking by imposters using digital photographs, video clips, printed pictures, or face masks.


  • Face and palm recognition multimodal standalone module
  • Built-in face & palm algorithms for full-cycle biometric recognition operations
  • Grayscale and true-colour RGB image capture in near-infrared (NIR)
  • Analysis of face and palm attributes
  • Built-in deep learning-based algorithms for high-speed enrollment and authentication
  • Highly secure liveness detection with anti-spoofing protection

Technical Specs:

  • Visible light camera with 256 grayscale levels
  • Optical wavelength: 850 nm
  • Illumination range: 0.01 lux to 20,000 lux
  • Palm tracking AE- exposure mode
  • Default output format: nv12
  • Recognition distance: 15cm ~ 40cm / 6 inches ~ 15 inch
  • Posture adaptability: yaw≤20°, pitch≤20°, roll≤90°, bend≤15°
  • Accuracy: tar=98.2%@far=0.05%
  • Recognition mode: 1:1, 1:n
  • Template extraction time: < 300ms for palm and < 200 ms for face
  • Matching time: < 150 ms for palm and < 70 ms
  • Liveness detection