New method for singularity detection based on multilevel partitions in fingerprint images
一种基于多级分割的指纹奇异点检测新方法

Based on multilevel partitions in a fingerprint image,a new method for singularity detection in a fingerprint image was presented to improve the accuracy and reliability of the detected singularities.Firstly,singularities were detected based on the information of the orientation field for fingerprint image,which is estimated by different blocking sizes and different methods of orientation field estimation(smoothing or not smoothing).Secondly,based on the corresponding relationship of the singularities detected with multilevel blocking sizes and different methods of orientation field estimation,the singularities were confirmed accurately and reliably.Finally,the experiments were done in the NJU-2000 fingerprint database that has 2500 fingerprints,the results show that the method performs well and it is robust to poor quality images.