In a non-face-to-face identity verification environment, attempts at fraudulent authentication using fake faces are considered a major threat. For this reason, the performance of fake detection that distinguishes between captured facial images and videos and real faces is of utmost importance.

Alchera, a company specializing in image recognition artificial intelligence (AI), announced on the 8th that it had received certification for its AI performance in the V&V (Verification & Validation) evaluation from the Telecommunications Technology Association (TTA) through its core AI facial recognition technology, FACE TRUST.

Alchera's core technology, Face Trust, uses AI to extract and compare user facial features to determine whether their identity is forged. It is mainly used in non-face-to-face identity verification processes using mobile and PC, and major banks, securities firms, and insurance companies are actually using Alchera's technology.

In particular, in the financial sector, when a customer attempts to authenticate themselves using an ID card, the 'other person acceptance rate', which is the rate at which others are judged to be the same person, and the 'self-rejection rate', which is the rate at which the same person is judged to be another person, are the most commonly used in evaluating facial recognition performance.

In this facial recognition evaluation, Alchera recorded a false rejection rate (FRR) of 4.09%, a false acceptance rate (FMR or FAR) of 0.00001%, and a false registration rate (FTE) of 0% based on ID images. Alchera evaluated that this simultaneously lowered the Foreign Acceptance Rate (FMR or FAR) and the Self-Rejection Rate (FRR), which are trade-offs in terms of security and usability.

In this forgery detection evaluation, Alchera recorded a forgery acceptance rate (SAR or APCER) of 0.03%, a real rejection rate (LRR or BPCER) of 3.33%, and a failure to acquire rate (FTA) of 0% by dealing with difficult facial images.

Since last year, according to the financial authorities' 'Financial Sector Voice Phishing Response Plan', the introduction of a facial recognition system has been recommended when opening a non-face-to-face account, and in particular, the importance of facial forgery detection technology is being emphasized.

Alchera is the first company in Korea to pass the 'Face Forgery and Modulation Detection Performance Test' of the international standardization organization iBeta without being aware of it, and recently won the paper award for 'Face Forgery Detection Based on Meta-Prompt Engineering Using Multi-Modal Learning', which combines computer vision and natural language processing AI, at IPIU 2024, the best image recognition AI academic conference in Korea, and is leading the way in technological advancement.

“Facial recognition and counterfeit detection technologies will become core technologies in all areas where identity verification is required,” said Hwang Young-gyu, CEO of Alchera. “Alchera’s technology guarantees accuracy and user convenience, and we will aggressively expand our business based on these strengths.”