改良式流向法之多級比對演算法指紋辨識晶片設計

Fingerprint Recognition Chip Design via the Fast Minutiae Extraction and Multi-level Verification Algorithm

黃一哲(Yi-Che Huang)

指導教授 ： 黃聰亮

淡江大學/工學院/電機工程學系碩士班/碩士(2007年)

https://doi.org/10.6846/TKU.2007.00771

摘要

在當今的社會中，資料安全的防護與門禁管理系統以及身分辨識是一項相當重要的課題。現今生物辨識技術逐漸廣泛的被運用在系統保密性以及身分辨別上，現在主要的生物辨識技術有，聲音、容貌、虹膜、指紋，其中指紋因為具有唯一性和不變性且具有相當高的可靠度而最常為人們所採用。
大部分的指紋辨識方法，都是先將指紋影像經過正規化、濾波雜訊、取指紋流向、流向濾波、細線化、偵測奇異點與特徵點。其中，又主要以指紋的奇異點(核心點、三角點)及紋線的特徵點(端點、叉點)的分布位置、數目作為之後比對的依據。指紋流向場若能精確的求得，那要擷取出特徵點以及核心點將是一件容易的事情。本文將提出一種新的改良式二值化法來取代以往大家常使用的Gabor濾波二值化法來加速指紋比對的速度以及成功率。
本論文又提出一種新的比對流程來加速比對單元的效率，在經過前面有效的影像處理步驟之後，我們省略了細線化直接做特徵擷取，並經由其核心點附近的流向對其做快速的指紋分類建檔，在比對單元方面，會先經由其分類建檔的資料庫來比對，將會大大的改善整個比對單元的辨識效率。

關鍵字

指紋 ； 辨識 ； 比對

並列摘要

Nowadays, the importance of the information protection and the entrance guard control system are quite important lessons. Today, biometric verification technique is applied on enhancing the system of maintaining secrecy and recognized identifications. Biometric verification techniques which are available now are mainly as follows: voice, appearance, iris, fingerprint. Fingerprint is so authentic adopted for people constantly because of its singularity and constancy.
Most identification methods would including normalization, noise filter, calculate fingerprint flow direction, flow filter, thinning, detect the positions of the singular points and minutiae. The primal methods are depend on the locations and numbers of ridge endings and bifurcations. If we can get the fingerprint flow field exactly, it will be easy to find out the minutiae and singular points. In this thesis, we propose a novel improved binarization to replace gabor filter. This method can accelerate the speed of the calculation of fingerprint flow and help us gain the flow more accurately.
We also advance a new verification process to promote the efficiency of verification. After a series of valid image preprocessing procedure, we detect the minutiae without Thinning, and use the flow nearing the core point to classify the fingerprints speedily. As for the verification unit, the new verification process would verify the fingerprints of its classified database first, which will improve the verification efficiency of the whole verification unit.

並列關鍵字

Fingerprint ； Identification ； Verification

參考文獻

[28] 郭丁溢，以濾波比對方式之指紋辨識晶片設計，淡江大學電機所碩士論文，民國94年。

[29] 李建潁，快速特徵搜尋之指紋辨識晶片設計，淡江大學電機所碩士論文，民國94年。

[35] 李釋書，利用嵌入式軟核心處理器之指紋辨識晶片設計與實作，淡

[4] E. R. Henry, Classification and Uses of Finger Prints. Routledge, London, 1900.

[6] K. Karu and A. K. Jain, “Fingerprint Classification,” Pattern Recognition, vol. 18, no. 3, 1996, pp. 389-404.