行政院國家科學委員會專題研究計畫 成果報告
植基於指紋類別的生物特徵識別系統
研究成果報告(精簡版)
計 畫 類 別 : 個別型
計 畫 編 號 : NSC 99-2221-E-151-057-
執 行 期 間 : 99 年 08 月 01 日至 100 年 07 月 31 日
執 行 單 位 : 國立高雄應用科技大學光電與通訊工程研究所
計 畫 主 持 人 : 王敬文
計畫參與人員: 碩士班研究生-兼任助理人員:陳亮瑜
碩士班研究生-兼任助理人員:洪銘&;#26318;
報 告 附 件 : 出席國際會議研究心得報告及發表論文
處 理 方 式 : 本計畫涉及專利或其他智慧財產權,1 年後可公開查詢
中 華 民 國 100 年 09 月 06 日
行政院國家科學委員會補助專題研究計畫
成果
報告
植基於指紋類別的生物特徵識別系統
計畫類別:■個別型計畫 □整合型計畫
計畫編號:NSC
99-2221-E-151-057
執行期間: 99 年 8月 1日至 100年 7月 31 日
執行機構及系所:國立高雄應用科技大學
計畫主持人:王敬文
共同主持人:
計畫參與人員:陳亮瑜、洪銘曎
成果報告類型(依經費核定清單規定繳交):■精簡報告 □完整報
告
本計畫除繳交成果報告外,另須繳交以下出國心得報告:
□赴國外出差或研習心得報告
□赴大陸地區出差或研習心得報告
■出席國際學術會議心得報告
□國際合作研究計畫國外研究報告
處理方式:
除列管計畫及下列情形者外,得立即公開查詢
■涉及專利或其他智慧財產權,□一年■二年後可公開查詢
中 華 民 國 100 年 8 月 8 日
Proceedings of the World Congress on Engineering 2011 Vol II WCE 2011, July 6 - 8, 2011, London, U.K.
目錄
報告內容
參考文獻
Proceedings of the World Congress on Engineering 2011 Vol II WCE 2011, July 6 - 8, 2011, London, U.K.
Evolutionary Optimization Approach for
Fingerprint Classification
Jing-Wein Wang, Member, IEEE
1
Abstract—To test the effectiveness of GHM multiwavelets in fingerprint classification with respect to scalar Daubechies wavelets, we study the evolutionary-based algorithm to evaluate the performance of each subset of selected feature. Comparatively studies suggest that the former transform features apparently contain more fingerprint information for discrimination than the latter.
Index Terms—GHM multiwavelt, Daubechies wavelet,
fingerprint classification, evolutionary-based algorithm
I. INTRODUCTION
ULTIWAVELETS have recently attracted a lot of theoretical attention and provided a good indication of a potential impact on signal processing [1]. In this paper, a novel fingerprint classification
scheme is proposed both to extend the
experimentation made in [1] and to test the effectiveness of the Geronimo-Hardin-Massopust (GHM) discrete multiwavelet transform (DMWT) [2] with respect to the scalar Daubechies wavelet [3]. Moreover, a point in genetic wavelet fingerprint analysis is that the chromosomes interact only with the fitness function, but not with each other. This method precludes the evolution of collective solutions to problems, which can be very powerful [4]. We further present an evolutionary framework for feature selection in which successive generations adaptively develop behavior in accordance with their natural needs. In the following sections, we give details of the propose fingerprint classification approach. The performance of the proposed method has been validated through experiments on the NIST special fingerprint database 4 (NIST-4) [5].
In the following sections, we give details of the propose fingerprint classification approach. Section II presents the example transformations for a fingerprint
image. Section III describes the proposed
coevolutionary feature selection scheme for
classification. In section IV, we present our
11111 Manuscript received Mar 06, 2011; revised Apr 07, 2011. This
work was supported in part by Taiwan, ROC, National Science Council under Grant NSC 99-2221-E-151 -057.
J. W. Wang is with the Institute of Photonics and Communications, National Kaohsiung University of Applied Sciences, Kaohsiung 807, Taiwan, ROC (phone: +886-930-943-143; fax: +886-7-383-2771; e-mail: jwwang@ cc.kuas.edu.tw).
experimental results tested on the NIST-4. Section V concludes the paper.
II. DISCRETE MULTIWAVELET TRANSFORMS
For a multiresolution analysis of multiplicity r > 1,
(MRA), an orthonormal compact support
multiwavelet system consists one multiscaling function vector Φ(x)(1(x),...,r(x))T and one
multiwavelet function vector
)) ( ..., ), ( ( ) (x 1 x x
Ψ r T. Both Φ and Ψsatisfy the
following two-scale relations:
) 2 ( 2 ) (x H Φ x k Φ k k
Z (1) ) 2 ( 2 ) (x GΦ x k Ψ k k
Z . (2)Note that multifilters {Hk} and {Gk} are finite
sequences of r × r matrices for each integer k. Let Vj,
j Z, be the closure of the linear span of
) 2 ( 2 /2 , ,j k j l jx k l , l = 1, 2,…, r. By exploiting
the properties of the MRA, as in the scalar case, any continuous-time signal f(x) V0 can be expanded as
) ( ) ( 1 0, , x k c x f r l k l k l
Z 2 (2 ) 1 2 / J, , x k c J r l k J l k l
Z 2 (2 ) 1 0 2 / , , x k d j r l J j k j l k j l
Z , (3) where
c jk cr jk
T k j, 1, , ,..., , , c (4)
d jk dr jk
T k j, 1, , ,..., ,, d (5) and dx k x x f cl,j,k
( )2j/2l(2j ) (6) dx k x x f dl,j ,k
( )2j/2l(2j ) . (7)M
Proceedings of the World Congress on Engineering 2011 Vol II WCE 2011, July 6 - 8, 2011, London, U.K.
For the two-dimensional discrete multiwavelet
transform, a 2-D MRA of multiplicity N for 2( 2)
R L
can be obtained by using the tensor product of two MRA’s of multiplicity N of L2(R). Fig. 1 shows a
fingerprint image of size 512512 and its one-level decomposition with the D4 wavelet transform and the GHM multiwavelet transform, respectively.
III. FEATURE SELECTION ALGORITHM
In the proposed method that is derived from the principles of the natural species evolution theory [6], individuals grouped in populations and thereafter
referred to as inter population Pb and intra
population Pw are randomly created. The two
populations have interdependent evolutions
(coevolution). The term inter reflects the reluctance of this individual for the opposite class. This reluctance is quantified by the mean square distance between pattern points that belong to different classes. An individual of the populationPb,Ix, will compete
with each individual of the population kernel Kb
which is the collection of individuals with best inter distances. The term Inter is formulated as follows:
Inter
(I Im) m x P Ix b
with ImKb, m = 1,…, M, , if , if ) ( D D D D D D I I m b x b m b x b m b x b x m p (8)where Db is the Euclidean distance between classes and p is a penalty. Conversely, the term Intra reflects the attraction of this individual for its own class. An individual of populationPw, Ix, will compete with each individual of the population kernel Kw which is the collection of individuals with best intra distances. A best individual of the population kernel Kb will compete with each of the best individuals of the opposite population kernel Kw. The combined results of these competitions directly provide the fitness function, and therefore the fitness function is
defined as a number composed of two terms:
= ( 1 - . ).( [Inter] – [Intra] ), (9) where is the weighting constant greater or equal to
one, is the number of features selected, is the number of training samples. The evaluation process of
is randomly combined with the Inter individual of
the population kernelKb and the Intra individual of the population kernelKw.
After computation of the fitness function for all the combination of the two kernel individuals, a feature selection step is activated for choosing the individuals allowed reproducing at the next generation. The strategy of feature selection involves selecting the best subset q,
q={u u1,...,q;u} (10)
from an original feature set ,
={
v v1 ,...,Q}, Q > q. (11) In other words, the combination of q features fromq will maximize equation (9) with respect to any
other combination of q features taken from Q, respectively. The new feature v is chosen as the (+1)st feature if it yields Max v Maxu Δ[Inter](u,v), (12) where
u , v , andΔ [Inter](u,v) = [Inter](u,v) [Inter](u) .
) (u
Inter]
[ is the evaluation value of equation (8)
while the feature u is selected and [Inter](u,v)is
the evaluation value of equation (8) while the candidate v is added to the already selected feature
u. In a similar way, the feature selection mechanism
minimizes intra measure and helps to facilitate classification by removing redundant features that may impede recognition. The proposed schemes consider both the accuracy of classification and the cost of performing classification.
To speed up such a selection process, we present a
packet-tree selection scheme that is based on fitness
value of equation (9) to locate dominant wavelet subbands. Following this innovative idea, the decomposed subbands at the current level, which can be viewed as the parent and children nodes in a tree, will be selected only if the predecessor at the previous level was selected. Otherwise, the scheme skips the successors and considers the next subbands. For each textured fingerprint, a representative tree by averaging the selected feature vectors over all the training samples is generated.
IV.GENETIC OPERATIONS
With a direct encoding scheme, the genetic representation is used to evolve potential solutions under a set of five-class 512 × 512 images with 256 gray levels (see Fig. 2) found in the NIST-4 database.
Proceedings of the World Congress on Engineering 2011 Vol II WCE 2011, July 6 - 8, 2011, London, U.K.
According to the roulette wheel selection strategy [7],
the combination of populations Pb and Pw
individuals with higher fitness value in equation (9) will survive more at the next generation. The combinative individuals selected in the previous step are used to as the parent individuals and then their chromosomes are combined by the following proposed combinative crossover criterion so as to toward the chromosomes of two offspring individuals. If the i-th genes of the inter and intra individuals are the same, then the i-th gene of the offspring individual is set as either individual. If not, the i-th gene of the offspring individual will be set as either individual at random. The size of each of the population remains constant during evolution. The mutation operation randomly changes a bit of the chromosome.
V. EXPERIMENTS RESULTS AND DISCUSSIONS The reported results have the following parameter settings: population size = 20, number of generation = 1000, and the probability of crossover = 0.5. A mutation probability value starts with a value of 0.9 and then varied as a step function of the number of iterations until it reaches a value of 0.01. Due to the curse of dimensionality, one hundred 256 × 256 overlapping subimages each class as training samples are used for the D4 wavelet and one thousand samples are used for the GHM multiwavelet. Textural features are given by the extrema number of wavelet coefficients [8], which can be used as a measure of coarseness of the fingerprint at multiple resolutions. Then, fingerprint classifications with feature selection were performed using the simplified Mahalanobis distance measure [9] to discriminate fingerprint textures and to optimize classification by searching for near-optimal feature subsets. The mean and variance of the decomposed subbands are calculated with the leave-one-out algorithm [9] in classification.
The performance of the classifier was evaluated with three different randomly chosen training and test sets. Algorithms based on the two types of wavelets have been shown to work well in fingerprint discrimination. The classification errors in Tables 1 and 2 mostly decrease when the used features are selectively removed from all the features at the decomposed levels 4 and 3, respectively. This decrease is due to the fact that less parameter used in place of the true value of the class conditional probability density functions need to be estimated from the same number of samples. The smaller the number of the parameters that need to be estimated, the less severe the curse of dimensionality can become. In the meanwhile, we also noticed that the multiwavelet outperforms the scalar wavelet with the packet-tree feature selection. This is because the extracted features in the former are more discriminative than the latter and, therefore, the
selection of a subband for discrimination is not only
dependent on the wavelet bases, wavelet
decompositions, and decomposed levels but also the fitness function.
To explore the performance of the proposed system, we report classification results on NIST-4 database with seven categories: right loop (437 images), left loop (484 images), tented (149 images), arch (530 images), S-type (twin loop) (110 images), whorl (241 images), and eddy (49 images). We compare our method to a few modern techniques as shown in Table 3. Our method not only achieves more accuracy in 4 and 5 classes than referred methods [10]-[13] with lower rejection rate, 7-class offers new report in classifying whorl, S-type, and eddy types, as well. On the other hand, we notice that there are some failures occurred in the experiments and the reasons can be summarized as the following. The indistinct ridges and valleys due to bad quality of the fingerprint image may lead to the fatal errors of wavelet extrema detection.
VI.CONCLUSIONS
This paper introduces a promising evolutionary algorithm approach for solving the fingerprint classification problem with the coevolving concept. While much of the researches are fighting to work out on the classification of four or five categories, even one or two seven classes, we have not joined the drive instead of reporting a new result on whorl, S-type, and eddy classes except general arch, tented arch, right and left loops.
REFERENCES
[1] V. Strela, P. N. Heller, G. Strang, P. Topiwala, and C. Heil, “The application of multiwavelet filter banks to image processing,” IEEE
Trans. Image Process., vol. 8, pp. 548-563, 1999.
[2] F. Keinert, Wavelets and Multiwavelets, Chapman & Hall, CRC, 2003.
[3] I. Daubechies, Ten lectures on wavelets. SIAM, Philadelphia, Penn., 1992.
[4] W. Siedlecki and J. Sklansky, “A note on genetic algorithm for large-scale feature selection,” Pattern Recognition Letters, vol. 10, pp. 335-347, Nov. 1989.
[5] C. I. Watson and C. L. Wilson, NIST special database 4,
fingerprint database. National Institute of Standards and Technology,
March 1992.
[6] T. Bäck, Evolutionary Algorithms in Theory and Practice:
Evolution Strategies, Evolutionary Programming, Genetic Algorithms, Oxford University Press, New York, 1996.
[7] D. E. Goldberg, Genetic algorithms in search, optimization, and
machine learning. MA: Addison-Wesley, 1989.
[8] S. Mukhopadhyay and A. P. Tiwari “Characterization of NDT signals: Reconstruction from wavelet transform maximum curvature representation,” Signal Processing, vol. 90, no. 1, pp. 261-268, 2010. [9] R. O. Duda, P. E. Hart, and David G. Stork, Pattern
Classification. Wiley-Interscience, 2000.
[10] A. K. Jain, S. Prabhaker, and L. Hong, “A multichannel approach to fingerprint classification,” IEEE Trans. on Pattern Recognition
and Machine Intell., vol. 21, no. 4, pp. 348-359, 1999.
[11] Y. Yao, G. L. Marcialis, M. Pontil, P. Frasconi, and F. Roli, “Combining flat and structured representation for fingerprint
Proceedings of the World Congress on Engineering 2011 Vol II WCE 2011, July 6 - 8, 2011, London, U.K.
classification with recursive neural networks and support vector machines,” Pattern Recognition, vol. 36, pp. 397-406, 2003. [12] Q. Zhang and H. Yan, “Fingerprint classification based on extraction and analysis of singularities and pseudo ridges,” Pattern
Recognition, vol. 37, pp. 2233-2243, 2004.
[13] J. Li, W. Y. Yau, and H. Wang, “Combining singular points and orientation image information for fingerprint classification,” Pattern
Recognition, vol. 41, pp. 353-366, 2008.
(a) (b)
Fig. 1. One-level decomposition for the NIST-4 fingerprint: (a) D4, (b) GHM.
Fig. 2. Fingerprint examples defined in Henry system: (a) right loop, (b) left loop, (c) tented arch, (d) arch, (e) whorl, (f) S-type (twin loop), (g) Eddy.
(a) (b) (c) (d)
Proceedings of the World Congress on Engineering 2011 Vol II WCE 2011, July 6 - 8, 2011, London, U.K.
TABLE 1
CLASSIFICATION RESULTS (CORRECT RATE IN %) USING THE D4 WAVELET PACKET DECOMPOSITION WITH COEVOLUTIONARY FEATURE SELECTION Sample Set = 1 = 2 = 3 = 4 = 5 1 90.47 90.48 90.29 90.49 90.43 2 90.41 90.38 90.62 90.29 90.49 3 90.41 90.43 90.38 90.51 90.49 Average 90.43 90.43 90.43 90.43 90.47 TABLE 2
CLASSIFICATION RESULTS (CORRECT RATE IN %) USING THE GHM MULTIWAVELET PACKET DECOMPOSITION WITH COEVOLUTIONARY FEATURE SELECTION Sample Set = 1 = 2 = 3 = 4 = 5 1 90.83 90.80 90.54 90.82 90.58 2 90.85 90.72 90.80 90.71 90.66 3 90.96 90.73 90.90 90.79 90.73 Average 90.88 90.75 90.75 90.77 90.79 TABLE 3
CLASSIFICATION RESULTS (CORRECT RATE IN %) COMPARED TO THE RELATED WORKS
Fingerprint
class Wang Jain et al.[10] Yao et al.[11]
Zhang and Yan [12] Li et al. [13] 7-class 90.88% (2.3%) * * * * 5-class 94.71% 90% (1.8%) 90% (1.8%) 84.3% 93.5% 4-class 95.36% 95.8% 94.7% 92.7% 95% * unavailable
Proceedings of the World Congress on Engineering 2011 Vol II WCE 2011, July 6 - 8, 2011, London, U.K.
計畫成果自評
本研究果成果與原計畫之目的完全相符合,由實驗結果得知我們使用的公開指
紋資料庫FVC DB1平均正確接受率為89.5%、而核心點的平均錯誤接受率則是
0.4%; DB2資料庫平均正確接受率為88%、而核心點的平均錯誤接受率則是
0.43%。我們也同時與其他文獻現有技術中最好的結果相比較,本研究之正確
接受率確實表現較佳。此外,本研究計畫之成果已發表於國外研討會ICCSE'11,
本研究相關技術已於2010年發表在SCI學術期刊長篇論文一篇。專利申請中一
件、已獲得一件,同時全國性競賽獲獎2項,人才培育3人。
1
國科會補助專題研究計畫項下出席國際學術會議心得報告
日期:100 年 8 月 8 日
一、參加會議經過
感謝國科會的補助,使我有機會到英國倫敦市(London)參加 2011 年計算機科學與工
程國際會議 ICCSE'11 (The 2011 International Conference of Computer Science and
Engineering)。此會議係屬於由 International Association of Engineers 所主辦 World
Congress on Engineering (WCE) 的 15 個研討會下其中一個研討會,主題涵括 Image
processing and computer vision、Pattern recognition、及 Artificial intelligence 等。各研
討會的 call-for-paper 由於主題明確,與會者所發表的研究主題與內容較易聚焦,比
較不會像一般的研討會報告主題內容差異較大。 WCE 近幾年舉行的地點並不相同,
但 大 都 以 英 國 或 美 國 境 內 為 主 。 這 次 會 議 是 選 在 位 於 London, Zone 1, South
計畫編號
NSC 99-2221-E-151-057
計畫名稱
植基於指紋類別的生物特徵識別系統
出國人員
姓名
王敬文
服務機構
及職稱
國立高雄應用科技大學
光電與通訊研究所/教授
會議時間
100 年 7 月 6 日至
100 年 7 月 8 日
會議地點
倫敦
會議名稱
計算機科學與工程國際研討會
ICCSE'11 (The 2011 International Conference of Computer Science and
Engineering)
發表論文
題目
進化式最佳演算法用於指紋分類
Evolutionary optimization approach for fingerprint classification
2
Kensington, Imperial College London 主校區,物理領域全球排名 14 (全歐洲排名 2) 的
物理大樓舉行,其學術研究成就相當優秀值得一覽。因本會議為大型的研討會,參
加的學者及學生代表來自許多國家,除歐盟會員國外、中東、部份非洲國家的人也
相當多,台灣的參加者如中央大學曾定章教授、崑山科大甘廣宙教授等人相對較少。
本人出席口頭報告並參與盛會,茲將參加會議經過與心得感想做個綜合性報告。
二、與會心得
此次會議自 0705-0708 共四天,首日為註冊及報到,第二~三日除 keynote 及 invited
speech 外為分組口頭發表與海報展示。0707 Steering Committee 亦同時召開會議檢討
本次大會之籌備與進行細節,做為明年舉辦 WCE2012 之參考。會議期間除了點心外
還供應午餐,解決了用餐上的問題。0707 的晚宴則在被定為二級古蹟的 170 Queen’s
Gate 餐廳用餐。當晚大家在具古典氣息的餐廳裝飾下用餐,並於活動尾聲中彼此互
道珍重、相約明年見。
由於 ICCSE'11 之主題涵概 Dimensionality reduction、 Feature extraction、Feature
Selection、及 Patter Recognition 等相關領域,與本人之研究範疇極為接近。況且,與
會者有些是以前在 papers 上曾見過名字的作者,親睹其本尊之演說另有一番感受,
與他們的交流更是獲益良多。此外,另一令我印象深刻的主題即是 Face recognition
研究現況。因本人目前正進行此主題之研究,趁此會議之便跟相關學者請益並討論
此技術未來之發展方向,收穫可謂豐富。最後,值得一提的是從 Imperial College
London 的資料看到,2009/10 該校年度的經費有將近 650 百萬英磅合台幣約 300 億,
該校人數 13,000 人左右約與本人服務之高雄應用科大相當,但本校每年教育部僅分
3
配 11 億加上其他各項計畫爭取也不過 20 億上下,比較之下令我覺得台灣的研究量
能實在差太多了!
三、建議
1. 圖訊識別在結合數位訊號處理技術,提升影像處理品質,及第三代通訊科技之
應用上均有很大的貢獻。一個專業若只有自我研究,不與其他團隊合作、不與國外
學者交流,其貢獻將是非常有限的,尤其在資訊科技高速發展的二十一世紀裡。因
此,如何拓展學術上的雙向交流,增廣見聞藉以提升研究品質,都是吾等迎接新一
代通訊科技應努力的方向。此次參加 ICCSE'11 國際研討會,特別覺得第三世界的國
家也正努力以赴的追趕,尤其是新興的大陸更是不在話下。相對的,台灣在歐洲地
區參與人數減少了,能見度也跟著降低了!
2. 建議學校應多鼓勵老師出國參加國際會議,以提昇教學與研究工作。同時建議
學校應可再提供適當的出國補助尤其是歐洲國家,以彌補國科會出國經費補助的不
足,亦可減輕老師出國的經濟壓力,相信必能增加老師出席國際研討會的意願。
三、攜回資料名稱及內容
1. 研討會論文集電子版 (Proceedings of the World Congress on Engineering 2011,
Editors: S. I. Ao, Len Gelman, David WL Hukins, Andrew Hunter, and A. M. Korsunsky,
Publisher: Newswood Limited, Organization: International Association of Engineers,
ISBN of Vol. II with pp. 920-1812: 978-988-19251-4-5)。
4
四、其他
1. 發表論文摘要如下(本論文入圍 ICCSE'11 最佳論文候選)
Abstract—To test the effectiveness of GHM multiwavelets in fingerprint classification
with respect to scalar Daubechies wavelets, we study the evolutionary-based algorithm to
evaluate the performance of each subset of selected feature. Comparatively studies suggest
that the former transform features apparently contain more fingerprint information for
discrimination than the latter.
2. 研討會網址 http://www.iaeng.org/WCE2011/ICCSE2011.html
3. 出席會議照片
4. 論文被接受發表之大會證明文件
Dear Dr. Jing-Wein Wang,
[Review result for the manuscript submissions in WCE 2011]
Thanks for your submission to the World Congress on Engineering 2011 (WCE 2011). It is our pleasure
to tell you that your manuscript
- paper number: ICCIIS_58
- title: Evolutionary Optimization Approach for Fingerprint Classification
has been accepted for the WCE 2011. Please read the attached review report. There is an appeal system
that we will arrange two other reviewers for your submitted paper if you do not agree with the grading or
comment(s) in this review.
5
For the accepted and registered papers, they will be included in the conference proceeding published by
IAENG (ISBN: 978-988-18210-6-5) in hardcopy. The length of each camera-ready paper will be limited
to 6 (IEEE style, double-column) pages. Each conference participant will be entitled to one proceeding at
the conference. The accepted papers will also be considered for publication in the special issues of the
IAENG journals and book chapters (soon after the conference). Revised and expanded version of the
selected papers may also be included as book chapters in the standalone edited books under the
cooperation between IAENG and publishers like America Institute of Physics, and Springer. The
proceeding will be indexed in major database indexes so that it can be assessed easily.
Important Dates:
Camera-Ready Papers Due & Registration Deadline (extended): 8 April, 2011
WCE 2011: 6-8 July, 2011
***
The details about the registration are available at:
http://www.iaeng.org/WCE2011/registration.html
***
The details about the camera-ready paper format and templates
are available at:
http://www.iaeng.org/WCE2011/publications.html
To ensure the correctness of the format, you are required to check your
revised paper against the checklist available for download at
http://www.iaeng.org/publication/download/Checklist_camera-ready_general.pdf
You are advised to send it immediately once it is ready instead of waiting
till the deadline. Usually, authors need to submit revised version more
than once before getting accepted as camera-ready version.
***
The details about the accommodation are available at:
http://www.iaeng.org/WCE2011/travel.html
***
More details about the WCE Best Paper Awards are available at:
http://www.iaeng.org/WCE2011/program.html
***
We are looking forward to your presentation of this paper and
would like to express our warm welcome to you in advance.
If you have any question, you are also welcome to tell us.
6
Best regards,
Anna Lee
Assistant Secretary
WCE 2011 Organizing Committee
http://www.iaeng.org/WCE2011/
Engineering Letters http://www.engineeringletters.com/
IAENG International Journal of Computer Science http://www.iaeng.org/IJCS/
IAENG International Journal of Applied Mathematics http://www.iaeng.org/IJAM/
IAENG Secretariat
Unit 1, 1/F, 37-39 Hung To Road, Hong Kong
The World Congress on Engineering 2011 (WCE2011)
Paper No.: ICCIIS_58 (The 2011 International Conference of Computational Intelligence and Intelligent
Systems)
Paper Title: Evolutionary Optimization Approach for Fingerprint Classification
Ratings: 5=excellent, 4=good, 3=average, 2=poor, 1= unacceptable
1. Content
Technical quality (1-5) : [ 4 ]
Technical originality (1-5): [ 4 ]
2. Presentation
Overall format (1-5): [ 4 ]
Abstract (1-5) : [ 4 ]
English (1-5) : [ 4 ]
3. Recommendation (tick one)
Accept : [ yes ]
Accept with minor revision: [ ]
Accept with major revision: [ ]
Reject : [ ]
Recommendation for Best Paper Awards competition : [ yes ]
Recommendation for the edited book : [ yes ]
4. Brief summary and further comments for the author(s) if any for improvement of the paper.
(Use a separate sheet if necessary)
7
The authors studied the evolutionary-based algorithm to evaluate the performance of each subset of
selected feature in fingerprint classification. The results are interesting.
國科會補助計畫衍生研發成果推廣資料表
日期:2011/08/10國科會補助計畫
計畫名稱: 植基於指紋類別的生物特徵識別系統 計畫主持人: 王敬文 計畫編號: 99-2221-E-151-057- 學門領域: 圖形辨識研發成果名稱
(中文) 利用奇異值分解於指紋影像之增強及切割方法及其系統(英文) Method of Using Singular Value Decomposition for Enhancing and Segmenting
Fingerprint Images and a System Thereof
成果歸屬機構
國立高雄應用科技大學發明人
(創作人)
王敬文,邱惠琪技術說明
(中文) 一種利用奇異值分解於指紋影像之增強及切割方法包含:利用一奇異值分解法分 解一原始指紋影像,以獲得一增益指紋影像;將該增益指紋影像進行能量轉換, 以獲得一能量分佈圖;及利用該能量分佈圖尋找指紋輪廓,以獲得數個標界,其 圍繞形成一切割多邊形。該指紋影像之增強及切割系統包含一輸入單元、一演算 單元及一輸出單元。該輸入單元用以輸入該原始指紋影像,該演算單元用以產生 該標界及切割多邊形,而輸出單元用以依該切割多邊形輸出該切割指紋邊界影像。(英文) A method of using singular value decomposition for fingerprint images includes:
decomposing an original image in a singular value decomposition manner to obtain an enhanced image; transforming energy of the enhance image to obtain an energy distribution; searching a fingerprint boundary by the energy distribution to obtain a plurality of landmarks to surround a segment boundary polygon. The enhancing and segmenting fingerprint image system includes an input unit, a calculating unit and an output unit. The input unit is used to input the original image, the calculating unit is used to generate the landmarks and segment boundary polygon, and the output unit is used to output a segmented fingerprint image according to the segment boundary polygon.