指紋特徵點定位研究

A study of point location of Fingerprint feature

黃怡昀(Yi-Yun Huang)

指導教授 ： 孫天佑

中原大學/理學院/應用數學研究所/碩士(2019年)

https://doi.org/10.6840/cycu201900242

摘要

由於指紋的獨特性可被使用於識別身分，訪問管制和刑事調查是主要使用的兩個領域。在指紋分析的過程中，透過特徵點位置進行指紋比對以確認是否為同類指紋，因此特徵點的定位準確度十分重要。在Henry及FBI對特徵點位置的定義難以透過計算機程序完成，必須依靠人類專家進行定位，且定位方法較為主觀，而Liu使用深度神經網絡提出了一種計算式定義特徵點的方法，可在像素級定義特徵點並能在非常小的圓中有效地定義特徵點。
在本論文中，將依照 Liu在2018年所發表的”Fingerprint Analysis and Singular Point Definition by Deep Neural Network”中的方法進行驗證所提出計算式定義核心點位置的演算法正確性，是否能夠定位於實際核心點位置。
根據結果顯示，此演算法在箕形紋類型的指紋上容易出現定位偏移的問題，此外，在此論文中也提出了校正偏移定位的想法。

關鍵字

特徵點定位 ； 指紋 ； 深度神經網路

並列摘要

The unique nature of fingerprints can be used to identify identity. Access control and criminal investigation are two major fields that take advantages of the unique nature of fingerprints. In the process of fingerprint analysis, fingerprint comparison through the position of singular points to confirm whether there are the same fingerprint. Therefore, the positioning accuracy of singular point is very important. The definition of the position of singular points in Henry and FBI are very hard to implement by computer programs. They must rely on human experts to locate, and the positioning method is more subjective than computational definitions. Liu uses a deep neural network to propose a the method of computational definition of singular point. The proposed method defines singular point at the pixel level and efficiently define singular point. in a very small circle
In this paper, the experiment verify correctness of the algorithm for defining the position of the singular point according to the method of "Fingerprint Analysis and Singular Point Definition by Deep Neural Network" published by Liu in 2018 , and determine if it can be located at the actual location of singular point .
The results suggested that the algorithm is prone to the problem of positioning offset on the fingerprint of loop type. In addition, the idea of correcting offset positioning has also been proposed in this paper.

並列關鍵字

location of singular point ； fingerprint ； deep neural network

參考文獻

[1] Bazen, A.M. and Gerez, S.H. (2002) Systematic methods for the computation of the directional fields and singular points of fingerprints’, IEEE Trans. on PAMI, 24(12), pp.905–919.

[2] Bo, J., Ping, T. and Lan, X. (2008) Fingerprint singular point detection algorithm by Poincaré index, WSEAS Trans. on System, 7(12), pp.1453–1462.

[3] Cappelli, R. and Ferrara, M., (2012) A fingerprint retrieval system based on level-1 and level-2 features, Exp. Syst. Appl. 39, pp.10465–10478.

[4] Dass, S.C. (2004) ‘Markov random field models for directional field and singularity extraction in fingerprint images’, IEEE Trans. on Im. Proc., 13(10), pp.1358–1367.

[5] Federal Bureau of Investigation (1984) The Science of Fingerprints: Classification and Uses, U.S. Government Printing Office, Washington, D.C.