摘要
本文提出了一种改进的指纹纹线方向场计算方法,其中重点分析了奇异点模型中旋转角度的估计方法,并在复数域内简化了纹线方向场的优化计算.实验结果显示该方法充分利用了指纹纹线方向在全局上的渐变性和在奇异点附近的独特性,可以有效降低指纹图像中局部噪声的影响,提高指纹纹线方向场的计算精度,改进指纹图像的增强效果.
This paper introduces an improved algorithm for fingerprint direction field estimation, the rotation angle estimation during constructing the singular point models is analyzed in details. The fingerprint direction field optimization is also simplified in complex field with nearly the same results. Experiments show that our algorithm, which take advantage of the typical ridge patterns near the singular points （cores and deltas） and the orientation character that the ridge curvature changes coherent in most regions of fingerprint images, can suppress the disturbing of different kind of noises in fingerprint images, improve the accuracy of the ridge direction estimation and fingerprint enhancement results.
出处
《小型微型计算机系统》
CSCD
北大核心
2006年第11期2047-2051,共5页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(60403044
60373070)资助.
关键词
指纹
方向场
梯度
优化
奇异点
多项式模型
fingerprint
direction field
gradient
optimization
singular point
polynomial model
