A SVM-based method for the estimation of fingerprint and palmprint orientation

A SVM-based method is proposed in this paper, which is the first time that SVM is applied to the estimation of fingerprint and palmprint orientation, to the best of our knowledge. The orientation estimation problem can be regarded as a multi-classification problem by dividing the range [0 180] into 8 segments uniformly, which allows the introduction of classification methods and does not reduce the performance. In addition to the direct presentation of the estimated orientation, the rates of accuracy and recall of the minutiae are considered as statistical evaluation indexes of the performance to compare our method with the gradient-based method. Detailed analytical and experimental results are given to verify the feasibility and efficiency of our method.