Acquisition and extraction of surface and internal fingerprints from optical coherence tomography through 3D fully convolutional network

Abstract

Optical coherence tomography (OCT) is a non-destructive, high-resolution, and non-invasive imaging technique that has recently been applied to biometric field for three dimensional (3D) fingertip data acquisition. Fingertip volume data contains the epidermis and part of the dermis layers representing the surface fingerprint and internal fingerprint, respectively. As a 3D and under skin feature, internal fingerprint provides much higher anti-counterfeiting capability compared with traditional 2D fingerprint. Furthermore, it is not affected by fingertip conditions such as wet, worn out (old) and small scars/injuries. However, the extraction of surface and internal fingerprints is challenging due to the speckle noise and the low contrast of the data. In this paper, a spectral domain OCT system is constructed for fingertip volume data acquisition. A simultaneous and automatic surface and internal fingerprint extraction method with 3D fully convolutional network is proposed. The 3D fully convolutional network is adapted for the contour extraction of stratum corneum and papillary junction with the inter-slice spatial dependency reserved by three-dimensional convolution. Subsequently, surface and internal fingerprints are extracted. Experiments are performed to demonstrate the effectiveness of our proposed method both qualitatively and quantitatively. The internal and external fingerprints are also compared and discussed.