摘要
提出了基于方向特性的指纹预处理算法,包括前景/背景提取、平滑、二值化、二值滤波、细化5部分。该算法通过计算指纹方向图对指纹图像进行分割,依据方向场实现二值化,最后采用骨架提取技术细化指纹。这套算法成功减少了原特征的丢失和伪特征的增加。实验结果显示,该算法能够去除噪声,得到较高质量的指纹图像。
A set of algorithms for fingerprint image preprocessing based on directional characteristic is presented, including foreground/background segmentation, smoothing in gray image, binarization, filtration in binary image and thinning. This algorithm partitions fingerprint image by calculating the directional image of the fingerprint, then achieves binarization with the directional fields, and finally adopts the technique of pick - u- ping framework to thin the fingerprint image. The preprocessing method can protect true minutia and avoid false minutia. The experimental results show that the method can eliminate the noise and get fingerprint image with high quality.
出处
《电子科技》
2007年第11期18-21,共4页
Electronic Science and Technology
关键词
指纹识别
预处理
方向特性
fingerprint recognition
preprocessing
directional characteristic
