회사 소개
EIWAT Co., Ltd. is an Osaka-based engineering company founded in 1972, with businesses spanning metal processing, renewable energy, smart community, cyber security, IoT and AI.
At K-Safety Expo 2026 we present SAFR, an AI facial recognition platform for which EIWAT builds the infrastructure, applications and system integration in Japan.
SAFR delivers 99.87% recognition accuracy at around 100 milliseconds per match, three to ten times faster than comparable engines.
Its liveness detection blocks spoofing by photo or video without special cameras or user actions such as blinking, and enhanced occlusion handling keeps accuracy high for masked faces.
The platform runs on existing IP cameras or on dedicated SAFR-SCAN and SAFR-CAMERA terminals, and its REST API and SDK allow integration with access control, kiosk and security systems.
EIWAT holds ISO 9001, 14001 and 27001 certification.
At K-Safety Expo 2026 we present SAFR, an AI facial recognition platform for which EIWAT builds the infrastructure, applications and system integration in Japan.
SAFR delivers 99.87% recognition accuracy at around 100 milliseconds per match, three to ten times faster than comparable engines.
Its liveness detection blocks spoofing by photo or video without special cameras or user actions such as blinking, and enhanced occlusion handling keeps accuracy high for masked faces.
The platform runs on existing IP cameras or on dedicated SAFR-SCAN and SAFR-CAMERA terminals, and its REST API and SDK allow integration with access control, kiosk and security systems.
EIWAT holds ISO 9001, 14001 and 27001 certification.
주요 품목
SAFR - AI facial recognition software (99.87% accuracy, approx. 100 ms response)
SAFR-SCAN / SAFR-CAMERA terminals, and facial recognition using existing IP cameras
System integration for access control, suspicious person detection and attribute analysis
SAFR-SCAN / SAFR-CAMERA terminals, and facial recognition using existing IP cameras
System integration for access control, suspicious person detection and attribute analysis
