摘要
作为一种重要的生物特征识别手段,自动指纹识别近年来得到了广泛的关注并取得很大的进展,指纹方向信息提取是自动指纹识别技术研究内容中一个非常重要的内容,但从低质量指纹图像中准确、可靠地提取方向信息却是一件很困难的事情。提出了一种多级分块尺寸下的指纹方向信息提取算法,依据大分块尺寸下所提取的方向信息对小分块尺寸下所提取的方向信息进行可靠性判断,然后进行相应的调整。实验结果证明,即使对于低质量指纹图像,该算法也可以快速、准确的提取指纹方向信息。
Automatic fingerprint identification system (APIS), as a kind of crucial means of biometrics, has been studied in recent years. Fingerprint orientation extraction is one of the significant widely contents in automatic fingerprint identification research. But it is very hard to extract orientation message from low-quality fingerprint images accurately. The paper proposes an algorithm for distilling fingerprint orientation in the multilevel block size, which adjusts the fingerprint orientation in the small block size resting on the fingerprint orientation in the big block size. The algorithm can extract fingerprint orientation accurately and quickly, even for low-quality fingerprint images. Experimental results indicate that the fingerprint orientation depicts tendency of transformation of fingerprint ridges.
出处
《南京大学学报（自然科学版）》
CAS
CSCD
北大核心
2003年第4期476-482,共7页
Journal of Nanjing University（Natural Science）
基金
南京大学重大应用研究预研项目基金(2001-03)
