Analysis and reconstruction of noisy finger vein infrared optical images

Biometric technology is required, along with technological advances in the digital world. In this digital age, authentication in digital transactions is necessary, especially for transaction security. We can utilize our unique body organs for the authentication process. One part of the human body that can be functioned for biometric is the finger vein. Finger vein can be used for biometric because every human living, finger vein pattern varies and unique for each person. Therefore, determine the physical characteristics of each person specifically. 780-980 nm near-infrared spectrum is effective for this application because of its portability and non-invasive trait. However, oftentimes, imaging results using infrared radiant contain noise. Notably, noise salt & pepper. This noise-reducing image information accuracy and making the extraction process more difficult, as a result, causing finger vein patterns difficult to recognize. To reduce the problem mentioned above, image noise must be reduced. One of the filters that can be used for reducing salt & pepper noise is the median filter. In the imaging transmitting mode, the intensity of infrared radiant absorbed into finger vein affecting imaging quality result. The bigger intensity, the bigger the transmittance radiance output unless saturation happens. As a result, the finger vein image becomes clearer. Moreover, attenuation intensity transmittance output value is influenced by wavelength. This process is simply according to Rayleigh scattering theory. Differentiation in wavelength, giving an effect in scattered intensity value, then affects attenuation intensity. Attenuation intensity produces different noises depends on the wavelength. This study suppresses the noise and analyzes the impact of wavelength variation on optical imaging process and noise level produced in 780-980 infrared spectrum span.