A Novel Mathematical Algorithm to Represent and Reconstruct some Bio-Metric Patterns

Fingerprint and Iris are significant biometric tools used for authentication. This paper proposes a novel methodology by which the biometric patterns can be visualized as a set of Bezier curves and hence represented by the corresponding Bezier points, resulting in considerable reduction in the file size. This scheme utilizes the Bezier curve representations for effective compression of all those images. Initially, the ridges and furrows present in the fingerprint and iris image are extracted along with their coordinate values. The control points are determined for all the ridges and furrows. The control points of all the ridges determined are stored and are used to represent the fingerprint and iris image. When needed, those can be reconstructed from the stored control points using Bezier curves. The quality of the reconstructed fingerprint and iris is also maintained. Thus the proposed scheme achieves considerable memory reduction in storing the biometric patterns.