Kneron announced that its Face Recognition Module v1.0 has recorded zero-error performance in independent testing against the ISO/IEC 30107 (Presentation Attack Detection) and ISO/IEC 19795 (Performance Testing) standards.
The month-long evaluation was performed by Fime, a laboratory accredited by NIST’s National Voluntary Laboratory Accreditation Program (NVLAP Lab Code 600365-0) for biometric testing and recognized by the FIDO Alliance for Biometric Component Certification.
Testing confirmed the following results:
ISO/IEC 30107 Presentation Attack Detection (PAD): 0.0% APCER (Attack Presentation Classification Error Rate) across all 12 tested attack vectors, including printed photos, curved and folded images, screen replay attacks, paper masks, complex multi-layer masks, full 3D-printed head models, and hood masks.
ISO/IEC 19795 Performance Testing: Across hundreds of thousands of match comparisons with 150 subjects, the module achieved 0.0% FAR (False Acceptance Rate), 0.0% FRR (False Rejection Rate), no capture failures (FTAR), and no enrollment failures (FTER). Performance remained stable under low light, strong light, and challenging indoor conditions.
The evaluation utilized a racially and demographically diverse dataset to represent global deployment scenarios, covering a range of racial backgrounds, ages, and facial characteristics.
The module operates fully on-device using Kneron’s ultra-low-power KL-series NPU architecture. It incorporates NIR (Near-Infrared) imaging and 3D sensing capabilities to distinguish real human faces from masks, replicas, or 2D spoofs. The system maintains millisecond-level response times even with glasses, facial hair, or difficult lighting conditions.
Albert Liu, founder and CEO of Kneron, stated that the zero-error performance under NVLAP-validated conditions marks a milestone for privacy-preserving edge AI. He noted that the results provide confidence for partners and customers to replace outdated, easily spoofed systems with next-generation authentication designed for current threats.
The certified module targets applications in smart door locks, home-entry systems, financial services, mobile payments, digital banking onboarding, identity verification, premium residential buildings, enterprise facilities, and access control environments. It enables on-device processing with no cloud uploads or external data handling, supporting battery-powered devices with extended operation.
Kneron provides certified reference designs to support partners in integrating the technology across the IoT ecosystem, aiming to reduce development timelines and technical barriers for secure, on-device AI deployment.
Detailed certification results and OEM reference designs are available at www.kneron.com/security.
