Face Recognition Web API Server FaceMe Platform

■ Details of Injection Attack Detection (IAD) < Protecting the Data Path from Image Capture to Face Recognition Engine > By monitoring the data path from the capture of facial images to the authentication process, we enhance the security of identity verification using smartphones and other devices by detecting attacks that inject counterfeit data by bypassing the camera module. < Protecting Each Layer of Applications, OS, and Camera > The injection attack detection of FaceMe SDK checks the status of each layer that constitutes identity verification and detects the following irregularities and risks: - Detection of spoofing using virtual cameras and camera API hooks - Integrity verification of applications and FaceMe SDK - Measures against tampering, debugging, and analysis of applications - Detection of rooted Android devices and jailbroken iOS devices - Detection of execution in emulators and virtual environments - Continuous monitoring of device status in the background

■ Details of the FaceMe Security 8.9 Update 1. Optimized Flow for Face Image Registration Administrators can now send notifications for face image registration to the individuals being registered when entering user information. The individuals can access the registration page using the URL or QR code provided in the notification from their smartphones, and follow the on-screen instructions to register their own face images. Key Features - Registration notifications via URL and QR code - Automated quality checks for face images - Support for approval and reshoot request flow by administrators - Support for bulk invitations via CSV 2. Support for Two-Factor Authentication with "Card + Face" The YouMe 500 facial recognition device from GIS now supports two-factor authentication (2FA) using "Card + Face." By combining card authentication with facial recognition, the risk of lending and fraudulent use, which is difficult to prevent with card-only systems, is reduced, enabling safer access control. This feature is suitable for use in environments requiring high security, such as access control to critical areas, factories, research facilities, and data centers.

**Improvement of Spoofing Prevention Accuracy and Performance** With FaceMe SDK 7.8, we enhance measures against increasingly sophisticated spoofing attacks by improving authentication accuracy and processing speed, thereby increasing the safety and reliability of facial recognition. - IR + RGB Spoofing Prevention: Detection accuracy improved by up to 35%, processing speed improved by up to 15% - 3D Spoofing Prevention: Detection accuracy improved by up to 56.8% for Intel RealSense D415 **Reducing Development Burden with the New Integrated API "FaceMe Authenticator"** We have integrated multiple processes that previously required individual implementation into a single API. Integrated Functions: - Face Detection + Feature Point Extraction + Spoofing Prevention + Authentication Processing Features: - Compatible with Windows / Ubuntu x64 / Android - 2D cameras, 3D depth cameras, and IR + RGB cameras can be used through a single interface - Significantly reduces development burden while maintaining high-accuracy authentication

In the latest "Age Estimation Evaluation Test" conducted by NIST, the AI facial recognition engine "FaceMe" ranked 5th in the world. 【What is the NIST FATE Age Estimation Evaluation Test?】 The NIST FATE Age Estimation Evaluation Test is one of the international benchmarks for facial analysis technology conducted by the National Institute of Standards and Technology (NIST) in the United States, which evaluates the age estimation accuracy of AI algorithms using a large-scale facial image dataset. 【Evaluation Results of FaceMe】 - World Ranking: 5th Among 20 vendors worldwide, FaceMe ranked 5th, demonstrating top-class accuracy in an international competitive environment. - Mean Absolute Error (MAE): 2.87 years FaceMe's average error in estimated age is 2.87 years, with only a 0.35-year difference from the top-ranked algorithm, achieving world-class accuracy. *MAE (Mean Absolute Error): A numerical representation of the error in estimated age (the smaller the number, the higher the accuracy). *NIST conducted the evaluation using 1 million facial images of individuals aged 18 to 30.