Panyayot Chaikan and Montri Karnjanadecha

pp. 71 - 79

Abstract

Our multimodal biometric system combines fingerprinting with a top-view finger image captured by a CCD camera
without user intervention. The greyscale image is preprocessed to enhance its edges, skin furrows, and the nail shape before
being manipulated by a bank of oriented filters. A square tessellation is applied to the filtered image to create a feature map,
called a NailCode, which is employed in Euclidean distance computations. The NailCode reduces system errors by 17.68%
in the verification mode, and by 6.82% in the identification mode.