) 3),6) 1) 3),7),8) 1) 3),9),10) DNA 1) 3)

Transcription

1 Summary We propose a person identification system using retinal fundus imagesthe proposed procedure for identification is based on comparison of an input fundus image with reference fundus images in the databasethis method is based on the measure of similarity defined as the cross-correlation coefficient calculated from the pixel valuesfive-hundred-eighty fundus images including eighty-eight same-person s image pairs were used for the estimation of the proposed technique. The FRR (false rejection rate) and the FAR (false acceptance rate) were found to be and , respectively. Thus this technique was useful for the biometrics. Key words: Biometrics, Fundus image, Region extraction, Blood vessel, Optic disc 1. Similarity Measurement Using Blood-Vessel Structure in Retinal Fundus Images for Biometrics Keisuke FUKUTA, Toshiaki NAKAGAWA, Yuji HATANAKA (Member), Yoshinori HAYASHI 4, Takeshi HARA,HiroshiFUJITA Department of Inteligent Image Information, Graduate School of Medicine, Gifu University Kowa Company., Ltd. Department of Electronic Systems Engineering, School of Engineering, The University of Shiga Prefecture 4TAK Co., Ltd. 1),2) 1)4), 1)3),5) 1)3),6) 1)3),7),8) 1)3),9),10) DNA 1)3) 1020

2 11),12) 1 13) 14)16) 14) 17) 18)20) 18) 19) 20) ) 3 abcde 1 2 Fig. 1 The structure of an eye Fig. 2 Flowchart of the overall process 1021

3 abc def 3 Fig. 3 The architecture of a fundus camera abc 4 Fig. 4 An example of retinal fundus image ) mm P- 21) 21) RGB (1) C = 1 J I {Ai, j ā}{bi, j b} IJ σ j=1 i=1 A σ B (1) AB I J ā b AB σ A σ B 1022

4 a b c 5 Fig. 5 Results of registration between input and reference images RGB ) Fig. 6 The image of blood-vessel area cut out from the original image in Fig. 4 and a calculation area used in similarity determination RGB 3. 2 A B 1023

5 A 156 B A: 78 B: 10 A 78 B A B JPEG RGB 24 bitsjpeg 1 40, t- p < FRRFalse Rejection Rate FARFalse Acceptance Rate 22),23) Table 1 The database of retinal fundus images used in the experiment Camera manufacturer Company K Company T Number of images Same person s pair Right eye Left eye Image size pixels A B 7 Fig. 7 Relative frequency of similarity between the same persons and different persons images in blood-vessel area 1024

6 8 FRRFAR Fig. 8 The relationship between False-Rejection RateFRRand False-Acceptance Rate FAR 9 Fig. 9 Relative frequency of similarity between the same persons and different persons images in original imagewhole area 40, FRR FAR Fig. 10 The relationship between FRR and FAR in original imagewhole area (a) 11 (b) 1025

7 a b Fig Distribution of correlation values caluculated in local regions a b = c = Fig Comparison of similarity between originalogwhole areaand blood-vesselbvarea images FAR FRR ID 24)28) 1026

8 ) Vol.44, No.1, pp.3 14 (2006). 2) hitech.pdf, (accessed April 23, 2007). 3) Vol.44, No.1, pp (2006). 4) Vol.44, No.1, pp (2006). 5) Vol.44, No.1, pp (2006). 6) Vol.44, No.1, pp (2006). 7) BM-ET300 Matsushita Technical Journal. Vol.49, No.5, pp (2003). 8) Vol.44, No.1, pp (2006). 9) Vol.44, No.1, pp (2006). 10) Vol.60, No.7, pp (2005). 11) C. Simon and I. Goldstein: A new scientific method of identification, New York State Journal of Medicine, Vol.35, No.18, pp (1935). 12) P. Tower: The fundus Oculi in Monozygotic twins: Report of six pairs of identical twins, Archives of Ophthalmology, Vol.54, pp (1995). 13) STEP (2004) 14) Eye-based biometrics, Biometric Technology Today, Vol.11, No.3 (2003). 15) Do the eyes have it?, Biometric Technology Today, Vol.9, No.10 (2001). 16) B.H. Robert: Biometrics: Personal Identification in Networked Society, Anil.Jain, Ruud Bolle, Sharath Pankanti, ed., Springer, New York (1999). 17) (2001) 18) RGB Symposium on Cryptography and Information Security (2007) 19) Symposium on Cryptography and Information Security (2007) 20) Symposium on Cryptography and Information Security (2007) 21) Vol.189-D, No.11, pp (2006). 22) Vol.44, No.1, pp (2006). 23) R. Cappelli, D. Maio, D. Maltoni, J.L. Wayman, and A.K. Jain: Perfomance evaluation of fingerprint verification system, IEEE Trans. on Pattern Analysis and Machine Intelligence, Vol.28, No.1, pp.3 18 (2006). 24) Vol.101, No.581, pp (2002). 25) J. Morishita, H. Watanabe, S. Katsuragawa, N. Oda, Y. Sukenobu, H. Okazaki, H. Nakata, and K. Doi: Investigation of misfiled cases in the PACS environment and a solution to prevent filing errors for chest radiographs, Academic Radilogy, Vol.12, No.1, pp (2005). 26) J. Morishita, S. Katsuragawa, K. Kondo, and K. Doi: An automated patient recognition method based on an image-matching technique using previous chest radiographs in the PACS environment, Medical Physics, Vol.28, No.6, pp (2001). 27) X Vol.101, No.581, pp (2002). 28) X Vol.59, No.10, pp (2003) CAD

9 CAD 1028