NEC Corporation has developed a multimodal biometric recognition technology for face and iris that can be widely applied for various purposes. With a compact system that only requires connecting a camera module to existing PCs or tablets, it can achieve high-speed and high-accuracy recognition for tens of millions of people, regardless of indoor or outdoor settings. NEC plans to advance the development and empirical research of this technology, focusing on applications in financial, retail, and entertainment sectors for payment and access control management, aiming for practical application by 2026.
Biometric recognition technology is becoming increasingly prevalent globally, especially in scenarios requiring strict identity verification such as finance and payments, where it is often combined with dual-factor authentication that includes showing cards and entering passwords. However, since cards and passwords carry risks of being forgotten, lost, or stolen, there is a need for technology that can securely authenticate identity without these methods.
NEC possesses the world’s most accurate face recognition and iris recognition technologies and launched a dedicated terminal in 2023 equipped with face recognition cameras, iris recognition cameras, processing CPUs, and displays for the face/iris multimodal biometric recognition solution.
This time, NEC prepared multiple sets of high-quality and low-quality images of the same person, and through machine learning of these images, successfully developed a technology that can infer iris features that could originally only be extracted from high-quality images, even from low-quality images (Figure 1). This technology is a world first.
Figure 1 Technical Overview (Image provided by: NEC)
Using this technology, even low-resolution images captured by face recognition cameras that contain a lot of noise can be used for iris recognition.
Additionally, through this technology, images captured by a single camera can achieve dual recognition of face and iris. Therefore, there is no need to adjust the camera position based on the user’s eye position, enabling high-speed recognition.
Moreover, the compact camera module can be integrated into devices such as POS machines and ATMs, and can also be carried on devices like tablets, thus promising to introduce face/iris multimodal biometric recognition technology into a wider variety of indoor and outdoor applications.
Original Article: “Science News” Translation: JST Objective Japan Editorial Department