提出了一种基于数学形态学与均值方差的复合指纹分割算法,改进了指纹图像分割效果,提高了指纹图像分割的精确度。对FVC2000、FVC2002和FVC2004数据库的实验结果表明,该算法能有效地对指纹图像进行自动分割,具有较快的速度和较好的鲁棒性。
In this paper, a composite image segmentation algorithm based on both mathematical morphology and mean-variance for fin- gerprint images has been proposed, and found to effectively improve the performance and accuracy of the fingerprint image segmenta- tion. The experimental results obtained on the Set B fingerprint databases of the FVC2000 FVC2002 and FVC2004 and our live-scanned fingerprint database validate this algorithm and prove a substantial improvement in the rapidity and robustness.