Development of a Spoof Face Detection System Utilizing Infrared Wavelengths Responsive to Human Skin

This paper presents the development of an access control terminal system capable of detecting spoof faces by utilizing infrared (IR) wavelength characteristics in the 920-960 nm band, which are responsive to human skin. The proposed system adjusts IR intensity according to installation distances of $80 \mathrm{~cm}, 100 \mathrm{~cm}$, and 120 cm to optimize detection performance. When IR light interacts with a human face, the system analyzes boundary features and saturation patterns between saturated and unsaturated facial regions. In contrast, spoof faces fabricated from materials such as paper, photographs, plaster, or silicone exhibit distinct saturation characteristics due to differences in material properties, despite maintaining similar facial shapes. The proposed method identifies spoof faces based on the unique property that IR light reflected from genuine human skin within the reference distance range does not saturate the facial image. For performance verification, spoof face samples were fabricated, and evaluation was conducted through the Telecommunications Technology Association (TTA). Experimental results demonstrate outstanding detection capability, achieving a 100% spoof face detection rate, thereby enhancing the overall security reliability of access control systems.