Singular Value Decomposition Based Fingerprint Gender Classification

Gender classification using fingerprint will be helpful in shortlisting the victims. In this paper, singular value decomposition (SVD) has been used for automatic fingerprint gender classification. The classification was achieved by extracting the spatial features of non-zero singular values obtained from the SVD of fingerprint images. The most robust K nearest neighbor (KNN) classifier has been used for gender classification. The evaluation of the system is carried out using internal database of 2200 fingerprints in which 1320 were male fingerprints and 880 were female fingerprints. The proposed method had produced an accurate identification of female gender to the maximum of 92.05% for the left hand little finger) and by average 83.30% for any other finger. Similarly, male gender is identified accurately to the maximum of 93.94% (for the right hand index finger) and an average 91.74% for other fingers. In addition, it was found that the success rate was higher for the male thumb fingers of male and reduces towards the small fingers. Similarly, the success rate in female small fingers is higher than the other fingers and reduces gradually towards the thumb finger. Detailed comparisons with earlier published results have been provided and our method offers better classification accuracy.