熊顯名 唐綺雯 張文濤
關鍵詞: 指靜脈; 特征提取; 灰度變化; 方向檢測; 四點法; 近紅外; 生物識別
中圖分類號: TN911.73?34; TP391.4 ? ? ? ? ? ? ? ? 文獻標識碼: A ? ? ? ? ? ? ? ? 文章編號: 1004?373X(2019)03?0061?04
Abstract: The current feature extraction method of figure vein has the problems of inaccurate line extraction, fractured characteristic texture, excessive noise and complicated extraction process. The near?infrared photoelectric transmissive vein acquisition method is used to collect the finger vein image. A four?point method is designed in the preprocessing stage to rapidly separate the vein region of the image from the background area. By analyzing the gray scale feature of the vein, the four?directional feature template detection operator perpendicular to the vein is designed to perform the feature extraction and experiment for the finger vein. The proposed algorithm can extract the feature line of vein quickly, and avoid the fracture problem of feature line, has simple extraction process and high efficiency, and the extraction speed of single vein line feature can reach up to 1.14 ms.
Keywords: finger vein; feature extraction; gray scale change; direction detection; four?point method; near infrared; biological recognition
手指靜脈作為人體內部生物特征,具有穩(wěn)定性和唯一性,無生命跡象手指因血液凝固無法檢測靜脈。指靜脈認證[1]作為一種非接觸活體識別具有更高的安全性和抗偽率,在近年來的生物識別研究[2]領域中引起廣泛重視。
由于指靜脈圖像采集裝置[3]的特殊性,使指靜脈圖像在采集過程中易受諸多因素的影響,如手指旋轉、移動、壓力、傳感器噪聲等,造成采集靜脈圖像質量低,靜脈特征提取[4?5]困難等問題。目前,常用的靜脈紋路提取與分割算法分為基于閾值圖像分割[6?7]和利用數(shù)學工具進行灰度分割[8]。文獻[9]利用閾值圖像法對手背靜脈圖像進行分割,此方法能夠對信噪比低、光照不均的圖像有較好的分割效果,缺點是容易丟失邊界信息,后期需進行嚴謹濾波除噪。文獻[10]提出重復線性跟蹤法,此方法能夠從不清晰靜脈圖像中提取出靜脈紋路,但提取過程復雜,不適用于較細靜脈提取。
本文提出一種基于灰度變化方向的特征檢測方法對手指靜脈紋路進行提取,設計四點法對圖像靜脈區(qū)域與背景進行快速分離,利用灰度變化方向特征設計四方向模板算子對分離后圖像進行靜脈紋路提取,提取速度快且能有效克服紋路斷裂及圖像質量低帶來的無法提取或提取結果不理想等問題。
圖8a)~圖8d)分別為本文選取的四個手指樣本用本文預處理方法得到的指靜脈預處理圖,圖8e)~圖8h)為以上四個樣本對應使用本文算法提取靜脈特征的效果圖。由圖像可以看出,盡管預處理階段存在光照不均、靜脈分支不明顯現(xiàn)象,但利用本文算法依然能得出有效且準確的靜脈紋路。
本文提出一種基于灰度變化的方向特征檢測算法對手指靜脈紋路進行特征提取。首先設計了四點法對圖像的靜脈區(qū)域與背景區(qū)域進行快速分離,再利用灰度變化方向特征設計四方向模板檢測算子對分離后靜脈區(qū)域進行紋路分割與提取,提取速度快且能有效克服由于圖像質量低帶來的靜脈無法提取或提取結果不理想等問題,降低了紋路特征斷裂問題,提取單幅圖像速度為1.14 ms,優(yōu)于傳統(tǒng)靜脈特征提取算法。編寫了基于VS2010平臺+OpenCV軟件操作平臺進行處理,驗證了算法提取靜脈特征的有效性。
參考文獻
[1] 秦斌.手靜脈身份識別技術[J].現(xiàn)代電子技術,2011,34(4):169?174.
QIN Bin. Technique of hand vein identification [J]. Modern electronics technique, 2011, 34(4): 169?174.
[2] 寒冰,李彬.生物特征識別技術的應用與發(fā)展新趨勢[J].中國安防,2010(8):40?43.
HAN Bing, LI Bin. Application and development trend of biometrics [J]. China security, 2010(8): 40?43.
[3] 黃建元,趙新榮,張長順,等.基于CMOS成像器件的手指靜脈圖像采集方法及裝置[J].紅外技術,2009,31(1):51?56.
HUANG Jianyuan, ZHAO Xinrong, ZHANG Changshun, et al. A method and device for collecting finger vein image based on CMOS imaging device [J]. Infrared technology, 2009, 31(1): 51?56.
[4] LEE E C, LEE H C, PARK K R. Finger vein recognition using minutia?based alignment and local binary pattern?based feature extraction [J]. International journal of imaging systems and technology, 2009, 19(3): 179?186.
[5] 王璐,張文濤.人體手掌靜脈圖像采集系統(tǒng)研究[J].激光與紅外,2013,43(4):404?408.
WANG Lu, ZHANG Wentao. Study on the collection system of human palm vein image [J]. Laser and infrared, 2013, 43(4): 404?408.
[6] 張晴,林家駿.超像素和閾值分割相結合的顯著目標檢測算法[J].現(xiàn)代電子技術,2016,39(14):95?99.
ZHANG Qing, LIN Jiajun. Significant target detection algorithm based on superpixel and threshold segmentation [J]. Modern electronics technique, 2016, 39(14): 95?99.
[7] 林喜榮,莊波,蘇曉生,等.人體手背血管圖像的特征提取及匹配[J].清華大學學報(自然科學版),2003(2):164?167.
LIN Xirong, ZHUANG Bo, SU Xiaosheng, et al. Characteristics extraction and matching of human hand dorsal vascular images [J]. Journal of Tsinghua University (science and technology), 2003(2): 164?167.
[8] 張萬濤,李維國.利用曲率信息的圖像分割改進模型[J].中國圖象圖形學報,2011,16(8):1385?1392.
ZHANG Wantao, LI Weiguo. An improved image segmentation model using curvature information [J]. Journal of image and graphics, 2011,16(8): 1385?1392.
[9] 王科俊,丁宇航,莊大燕,等.手背靜脈圖像閾值分割[J].自動化技術與應用,2005(8):19?22.
WANG Kejun, DING Yuhang, ZHUANG Dayan, et al. Threshold segmentation of the dorsal hand vein [J]. Automation technology and applications, 2005(8): 19?22.
[10] MIURA N, NAGASAKA A, MIYATAKE T. Extraction of finger?vein patterns using maximum curvature points in image profiles [J]. IEICE transactions on information and systems, 2007, 90(8): 1185?1194.
[11] 黎蔚,李繼杰,陳家新,等.手掌靜脈圖像增強算法研究[J].微電子學與計算機,2010,27(7):237?241.
LI Wei, LI Jijie, CHEN Jiaxin, et al. Research on the algorithm of palm vein image enhancement [J]. Microelectronics and computers, 2010, 27(7): 237?241.