摘要
重点研究了二值化后指纹图像细化算法.首先简介了改进的OPTA细化算法.在总结上述算法的优缺点基础上,提出更有效的基于数学形态学的查表细化算法.实验表明后一种细化算法效率更高、处理后几乎没有毛刺、效果更理想,为后续的工作奠定了良好的基础.所进行的算法高效化研究对于提升识别的准确度具有非常明显的积极作用,在论文的最后通过实证手段进行了证明.同时,该文最后采用Visual C++编程实现指纹图像预处理过程中的所有算法,算法效率较高,从实际需求角度为我国的指纹图像识别的细化处理提供了必要的理论依据,实践证明,该文所研究的数学形态学视角下的指纹图像细化算法高效研究具有一定的现实意义,应用于指纹图像识别工作中将会取得较好的效果.
As an important part of the fingerprint recognition and the underlying data sources, the pre- processing of the fingerprint image has an important rolef its quality is good or bad for follow-up fingerprint recognition accuracy has a very big impact, especially on the fingerprint feature extraction, is of great im- portance. Refined processing as the pretreatment of the core link, for fingerprint image recognition nature has the extremely important practical significance. But the question image thinning algorithms have many deficiencies, in the process of actual application shows great limitations, causes our country the fingerprint image thinning recognition career progress is slow. Based on the above situation, this paper focuses on the two value of fingerprint image thinning algorithm. This paper introduces the improved OPTA thinning algo- rithm. At the conclusion of the above algorithm based on the merits, some more efficient algorithm based on mathematical morphology table look-up thinning algorithm. In this paper, the algorithm is efficient for pro- moting the recognition accuracy has very obvious positive effect, in the end proved by empirical means. At the same time, this paper finally using the Visual C＋ ＋ programming realization of fingerprint image pre- treatment process of all the algorithmg, the algorithm has higher efficiency, from the actual demand per- spective for our image of the fingerprint identification of thinning processing to provide the essential theory basis, practice has proved, this paper studies the mathematical morphology from the Perspective of the fin- gerprint image thinning algorithm and efficient research has a certain realistic meaning, applied to finger- print image recognition work will achieve good resfilts.
出处
《湘潭大学自然科学学报》
CAS
CSCD
北大核心
2012年第3期93-97,共5页
Natural Science Journal of Xiangtan University
基金
国家自然科学基金项目(10771043)
关键词
指纹识别预处理
数学形态学
细化
fingerprint recognition preprocessing
mathematical morphology
thinning
