Fingerprint orientation field reconstruction by weighted discrete cosine transform

Orientation field represents the topological structure of the interleaved ridge and valley flows in fingerprint images. Although a number of methods have been proposed for orientation estimation, reliable computation of orientation field is still a challenging problem due to the poor quality of some fingerprints. This paper proposes a method to reconstruct fingerprint orientation field by weighted discrete cosine transform (DCT). First, the DCT functions are used to build the basis atoms for linear representation of orientation field. Then, the DCT basis atoms of low and high orders are combined with the weights determined by singularity measurements for orientation reconstruction. The weighted DCT model is further extended for partial fingerprints to gradually and iteratively reconstruct the orientations in noisy or missing parts of fingerprints. The proposed method can perform well in smoothing out the noise while maintaining the orientation details in singular regions. Extensive experiments have been done to compare the proposed method with some existing methods on NIST and FVC fingerprint databases in terms of the reconstruction accuracy of orientation field, fingerprint indexing performance, and fingerprint recognition accuracy. Experimental results illustrate the effectiveness of the proposed method in reconstructing orientation fields, especially for poor quality and partial fingerprints.