以濾波比對方式之指紋辨識晶片設計

Chip Design of Filter-based Fingerprint Recognition Method

郭丁溢(Ting-Yi Kuo)

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(2005年)

https://doi.org/10.6846/TKU.2005.00548

摘要

在電腦視覺和圖形辨識領域的科學家及工程師們，正努力實現一套智慧型系統用於辨識生活中各種不同有用的資訊。在生物特徵中包括有指紋、虹膜、臉形、掌紋、聲紋、簽名等，由於指紋具有獨特性與不變性且基於成本等各種考量，使得指紋辨識在各種生物資訊中最廣為科學家及工程師們所研究、發展，將指紋辨識應用於有關安全性的系統上。指紋辨識中比對的方式有許多，但最廣為研究使用的為Minutiae-based，而利用Minutiae-based比對方式所需處理的動作，例如：影像增強、二值化、細線化等這些動作往往是很費時；特徵點數目的不同，特徵點的匹配和比對亦是需要不少的計算時間。指紋影像經細線化後容易產生誤判的特徵點，且比對的準確度往往受到指紋影像品質的影響。
另有學者提出不以特徵點做為指紋的比對特徵，而將指紋影像視為具方向性的紋理結構，利用指紋紋線資訊做為比對特徵數據。利用Gabor filterbank頻率與方向擷取出指紋影像中全域與區域資訊當作固定長度的特徵向量(Fingercode)，而比對的方式為計算兩指紋影像Fingercode的歐式距離，計算出它們的相似或相異程度，故其比對速度快。本論文提出改善Ridge Feature-based比對方式，處理有關指紋影像偏轉的問題，增加辨識準確性與正確性，以Nios-Stratix 發展板實現成硬體晶片，提供了一個指紋辨識發展平台，以實驗室購買的電容式指紋擷取器，擷取16位同學左拇指與右拇指的指紋影像，每隻拇指按壓兩次共64張指紋影像，整體辦識成功率為96.875%。

關鍵字

指紋 ； 指紋碼 ； 蓋伯濾波 ； 紋理 ； 指紋流向 ； 參考點

並列摘要

The scientists and engineers of the field in computer vision and pattern recognition, make great efforts to realize that a set of intelligent systems are being used for recognizable all kinds of useful information in life. The biological characteristics include fingerprint, iris, shape of face, palm, voiceprint, signing, etc. Because fingerprint has uniqueness and invariable and consideration of cost, fingerprint is most widely used for personal identification in security. Among fingerprint matching algorithms can be broadly classified as minutiae-based and ridge feature-based. The widely used minutiae-based needs to deal with image enhancement, binarization, thinning that are very time-consuming. It usually detects many of the false minutiae and quality of fingerprint is critical fact of accuracy.
A.K. Jain[20] thought fingerprint feature as oriented texture and proposed filter-based algorithm use a bank of Gabor filters to capture both local and global information in fingerprint as a compact fixed length Fingercode. The matching is to finding the Euclidean distance between these Fingercode, hence the matching is very fast.
This thesis proposes improving ridge feature-based deal with the problem about rotation and transfer of fingerprint, and increase the accuracy. I implement the fingerprint identification to the system on a programmable chip with Altera’s Nios Stratix development kit and use capacitive sensor to capture the fingerprint of left thumb and right thumb of 16 classmates. The accuracy of matching is 96.875%.

並列關鍵字

Fingerprints ； FingerCode ； Gabor filter ； Texture ； Ridge flow ； reference point

參考文獻

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

[4] K. Karu and A. K. Jain, “Fingerprint classification,” Pattern Recognition, vol. 29, No. 3, 1996, pp. 389-404.

[6] A. K. Jain and S. Minut, “Hierarchical kernel fitting for fingerprint classification and alignment,” Pattern Recognition, vol. 2, 2002, pp. 469-473.

[7] B. G. Sherlock, D. M. Monro, and K. Millard, “Fingerprint enhancement by directional Fourier filtering,” IEE Proc. Vis. Image Signal Processing, vol. 141, 1994, pp. 87-94.

[8] D. C. D. Hung, “Enhancement and feature purification of fingerprint image,” Pattern Recognition, vol. 26, 1993, pp. 1661-1671.