Professional Documents
Culture Documents
Abstract—Iris analysis studies the relationship between human made great achievements in recent years. For example, biomet-
health and changes in the anatomy of the iris. Apart from the fact ric methods based on the spatial pattern of the iris are considered
that iris recognition focuses on modeling the overall structure of to attain a very high accuracy, and therefore, there has been an
the iris, iris diagnosis emphasizes the detecting and analyzing of lo-
cal variations in the characteristics of irises. This paper focuses on explosion of interest in iris biometrics in recent years [3]. At
studying the geometrical structure changes in irises that are caused the same time, people may now diagnose appendicitis based on
by gastrointestinal diseases, and on measuring the observable detecting tongue color, external characteristics on the tongue
deformations in the geometrical structures of irises that are re- or blood vessels under the tongue [4]; recognize gastritis and
lated to roundness, diameter, and other geometric forms of the cholecystitis by collecting and automatically analyzing pulse
pupil and the collarette. Pupil- and collarette-based features are
defined and extracted. A series of experiments are implemented conditions [5]; and apply breath analyses in medicine and clin-
on our experimental pathological iris database, including manual ical pathology [6]. Increasingly, more attention and demand
clustering of both normal and pathological iris images, manual have been placed onto such noninvasive, real-time, harmless,
classification by nonspecialists, manual classification by individ- and practical biometric methods to address health issues by
uals with a medical background, classification ability verification both individuals and hospitals.
for the proposed features, and disease recognition by applying the
proposed features. The results prove the effectiveness and clini- The iris is the “colored ring of tissue around the pupil through
cal diagnostic significance of the proposed features and a reliable which light enters the interior of the eye” [3], [7]. The iris is an
recognition performance for automatic disease diagnosis. Our re- inner body organ, but also has externally visible characteristics.
search results offer a novel systematic perspective for iridology The fact that an iris contains a large amount of characteristic
studies and promote the progress of both theoretical and practical information in terms of complex and precise structural and tex-
work in iris diagnosis.
tural details makes iris recognition and iris diagnosis (iridology)
Index Terms—Collarette-based feature, disease recognition, iris two important directions in biometrics study.
geometric deformation feature, pupil-based feature. Biometric methods based on the spatial pattern of the iris are
believed to result in very high accuracies, which have led to an
explosion of interest in iris biometrics in recent years. The main
I. INTRODUCTION
idea of iris recognition is that the minute details of the iris tex-
IOMETRIC methods identify people or analyze the status
B of human health based on physical or behavioral charac-
teristics [1], [2]. Most biometric methods measure human body
ture can be randomly determined during the fetal development
of the eye [8], [9], and the color of the iris can change as the
amount of pigment in the iris increases during childhood [3].
surface characteristics, such as finger prints, palm prints, face, However, this argument undermines the fact that the iris pos-
irises, tongue, breath, etc. [2]. Computer-aided biometrics has sesses different kinds of variability, which produces special iris
features. In fact, in order to assure a satisfactory performance
in most applications, iris recognition systems demand a regular
re-enrollment of users. This implies that the iris structure and
Manuscript received February 5, 2012; revised July 9, 2012; accepted surface may be affected by various conditions and changes to
September 24, 2012. Date of current version February 4, 2013. This work was
supported in part by the Specialty Leader Support Foundation of Harbin under iris codes may be obvious enough to cause mismatchings [10].
Grant 2003AFLXJ006, in part by the Foundation of Heilongjiang Province for Iridology methods based on structural and textural changes in
Scholars Returned from Abroad under Grant LC04C17, in part by the Fun- the local iris areas are regarded as promising in health surveil-
damental Research Funds for the Central Universities HIT.NSRIF.2012047,
and in part by the National Natural Science Foundation of China under Grant lance and subhealth diagnosis. The main idea of iridology is that
61271345. the health status will be reflected on the iris due to two reasons.
L. Ma, N. Li, W. Zuo, and K. Wang are with the School of Computer Science First, the anatomical characteristics of the iris bring about
and Technology, Harbin Institute of Technology, Harbin 150001, China (e-mail:
malin_li@hit.edu.cn). great variability to the iris. There are two muscles which are
D. Zhang is with the Shenzhen Graduate School, Harbin Institute of Tech- called the dilator and the sphincter muscles that control the size
nology, Shenzhen 518055, China, and also with the Department of Computing, of the iris to adjust for the amount of light that enters the pupil.
Biometrics Research Centre, Hong Kong Polytechnic University, Kowloon,
Hong Kong (e-mail: csdzhang@comp.polyu.edu.hk). Iris tissue does not follow a perfect “rubber sheet” model when
Y. Cai is with the Fourth Affiliated Hospital, Harbin Medical University, the pupil dilates and contracts [10]. This variation of the iris
Harbin 150001, China. surface characteristics becomes a common problem in all iris
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org. recognition methods, because it greatly affects the quality of a
Digital Object Identifier 10.1109/TITB.2012.2222655 match between iris images [3]. Injuries and indirect influences
TABLE I
DATA DISTRIBUTION FOR NORMAL AND PATHOLOGICAL IRISES IN THIS PAPER
TABLE III
MANUAL CLASSIFICATION OF NORMAL AND ABNORMAL IRISES
(AVERAGE PERFORMANCE ON TEN RESPONDENTS)
TABLE V
MEAN VALUE OF L 1 UNDER SEVERAL CONDITIONS FOR ESTIMATING THE
DIAMETER OF THE PUPIL
TABLE VI
STATISTICAL ANALYSIS OF PUPIL FEATURES FOR NORMAL AND
PATHOLOGICAL IRISES
(x −x )•(x −x )
is given by ρk = arccos |(x (( kk ++ 33 )) mm oo dd KK −x kk )||(x (( kk −3 ) mod K k
−3 ) m o d K −x k )|
.
Obviously, a greater L3 value indicates increased worsening
and minimum diameters of the pupil can be determined; hence, in the degree of smoothness.
a larger L1 value is obtained. Table V shows our experimental By using the previous definition of the pathological features
results for choosing the best number of directions. We tried to of an iris pupil, we analyzed 150 normal irises and 142 irises in
estimate diameters for six cases: horizontal and vertical direc- patients with digestive pathology. The statistic results are given
tions, the 4 directions as indicated previously, and 8, 12, 16, and later. Obviously, the majority of the parameters of normal and
32 directions that are equally divided in a circle. We calculated pathological irises have good classification results.
the L1 value for all of the 142 pathological iris examples in Fig. 4 shows a histogram of the iris examples in our database
our database, and provided the mean value of L1 in the table. based on the L1 feature. The same phenomenon can be seen
It is apparent that more directions augment a mean L1 value, for the L3 feature. However, the L2 feature does not have equal
and could make the L1 feature more significant. However, the differentiating power, as shown in Table VI.
computational cost was also greatly augmented. The four direc-
tions are chosen for a good balance between performance and B. Collarette-Based Features
computational costs.
It was also observed that the area between the pupil boundary
2) Pupil Largeness Feature L2 : The pupil largeness feature
and collarette contains large variations. Normal collarettes form
is the average ratio of the pupil diameter to the iris diameter,
a jagged closed loop. Diseases may cause collarettes to expand
and describes the amount of pupil that occupies an eye. When
or reduce with geometric changes. The collarette boundary can
disease causes the tension in the inner edge of the iris to change,
be estimated through an ant colony optimization based image
either tightening or breaking, the pupil is reduced or enlarged.
segmentation algorithm (ACO-ISA) proposed in our former re-
As the result, the largeness of relative area of the pupil in the
search [17]. Similar to pupil pathological features, collarette
iris is changed and reflected in the pupil iris ratio.
features may be quantitatively defined and calculated for iris
Here, pupil iris diameter ratio L2 is defined as the average of
diagnostic applications. The collarette features proposed in this
the horizontal, vertical, and two directional ratios between the
paper are as follows.
pupil and the iris diameter
1) Roundness of Collarette Boundary A1 : A1 is defined as
1 Dhinner Dvinner Dlinner Drinner the ratio between the maximum and the minimum values of the
L2 = + + + (2) six diameter measurements of a collarette, as shown below:
4 Dhouter Dvouter Dlouter Drouter
where Dhouter , Dvouter , Dlouter , and Drouter are the iris widths in max Dhcollar , Dvcollar , Dlcollar , Drcollar , Dm collar
, Dncollar
A1 =
the horizontal, vertical, 45◦ , and 135◦ directions. min Dhcollar , Dvcollar , Dlcollar , Drcollar , Dm collar , D collar
n
3) Pupil Edge Roughness L3 : This neatness and smoothness (4)
feature in the inner edge of the iris is very important. It is the where Dhcollar , Dvcollar , Dlcollar , Drcollar , Dm
collar
, and Dncollar are
most important indicator of a healthy digestive system. Given the collarette diameters in the horizontal, vertical, 30◦ , 60◦ ,
an edge that contains points (x1 , x2 , . . . , xk , . . . , xK , x1 ), the 120◦ , and 150◦ directions. The reason that these six diameters
L3 roughness feature is defined as: are used for collarette features instead of the four that were used
for the pupil is that the collarette boundary is much rougher
1
K
than the pupil boundary, and 6 is a good balanced number of
L3 = |ρk − ρ̄|. (3) diameters in this case.
K
k =1
Clearly, standard circles have A1 = 1.0; and statistics for
Here, ρk is the curvature at element xk , and ρ̄ stands for the a normal person produce a collarette that has a nearly round
mean curvature. In our research, a third-order curvature is ap- boundary with 1.13 < A1 < 1.27.
plied and ρk ∈ [0, 2π] is measured as the angle between the line 2) Collarette Iris Diameter Ratio A2 : The anatomy of an
(x(k + 3) m o d K − xk ) and the line (x(k −3) m o d K − xk ), as what iris reveals that the collarette makes up 1/3 of the iris area. A2
228 IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, VOL. 17, NO. 1, JANUARY 2013
TABLE VII
STATISTIC RESULTS FOR COLLARETTE FEATURES
Fig. 7. Iris images from a healthy person and those suffering from stomach
diseases. (a) Gastric ulcer. (b) Gastroduodenal bleeding. (c) Gastric cancer.
(d) Normal.
V. CONCLUSION
In this paper, we have proposed a series of geometric features
to quantitatively analyze pathological changes in irises caused
by common illnesses, such as gastrointestinal diseases. A statis-
tical investigation is performed on a database with irises from
normal individuals and patients who suffer from gastrointestinal
diseases, and demonstrated well the diagnostic abilities of the
proposed features. We also proposed a distinctive computerized
Fig. 10. Gastrointestinal disease, enteritis system disease, and normal iris
iris diagnostic approach, and the experimental results are quite
distribution per the first two main features given by a PCA analysis based on promising.
pupil and collarette shapes. The main contribution of this research is that not only is the
TABLE VIII
mapping relationship between iris characteristics to Western
CLASSIFICATION PERFORMANCE ON TYPICAL DISEASES VERSUS medicine defined diseases revealed, but also contributes to both
HEALTHY PERSON iris diagnosis and recognition by separating the required iris
features into globally structural and locally geometrical features.
This will undoubtedly prompt the modernization of traditional
iris diagnosis, and more importantly, lessen the research gap
between iridology research and clinical applications.
REFERENCES
that the Fisher linear classifier method is simple, but gives a [1] D. Zhang, Automated Biometrics—Technologies and Systems. Norwell,
slightly lower recognition rate; the SVM method may be the MA: Kluwer, 2000, pp. 18–25.
[2] A. Jain, A. Ross, and S. Prabhakar, “An introduction to biometric recog-
best in recognition, but to take into account more features in the nition,” IEEE Trans. Circuit Syst. Video Technol., vol. 14, no. 1, pp. 4–20,
experiment requires the combination of prior knowledge and Jan. 2004.
more direction, so its recognition performance is lacking; and [3] K. W. Bowyer, K. Hollingsworth, and P. J. Flynn, “Image understand-
ing for iris biometrics: A survey,” Comput. Vis. Image Understanding,
finally, the SOM network method is also very effective, because vol. 110, pp. 281–307, 2008.
the number of neural network units can be further added into [4] B. Pang, D. Zhang, and K. Wang, “Tongue image analysis for appendicitis
the model to enhance its capacity, and also because no a priori diagnosis,” Inf. Sci., vol. 175, pp. 160–176, 2005.
[5] D. Zhang, L. Zhang, D. Zhang, and Y. Zheng, “Wavelet based analysis of
classification knowledge is needed for the exact classification of Doppler ultrasonic wrist-pulse signals,” in Proc. Int. Conf. Biomed. Eng.
an object. Therefore, we believe that the SOM network method and Inf., vol. 2, pp. 539–543, 2008.
may be more increasingly competitive in future practical use of [6] D. Guo, D. Zhang, N. Li, and L. Zhang, “A novel breath analysis system
based on electronic olfaction,” IEEE Trans. Biomed. Eng., vol. 57, no. 11,
iris diagnostic systems. pp. 2753–2762, Nov. 2010.
Table VIII shows the average recognition rate of the three [7] C. Oyster, The Human Eye Structure and Function. Sunderland, MA:
methods in this practical disease recognition application. All Sinauer Associates, 1999.
[8] J. Daugman and C. Downing, “Epigenetic randomness, complexity and
three methods achieve nearly a 90% recognition rate for the indi- singularity of human iris patterns,” Proc. Roy. Soc. London B, vol. 268,
viduals with the two diseases and the healthy individuals. These pp. 1737–1740, 2001.
good recognition rates also prove the clinical diagnosis signifi- [9] T. J. Buchanan, C. J. Sutherland, R. J. Strettle, T. J. Terrell, and A. Pewsey,
“An investigation of the relationship between anatomical features in the
cance and the effectiveness of the proposed pupil and collarette iris and systemic disease, with reference to iridology,” Complementary
features. These features allow a large degree of choices for clas- Therapies Med., vol. 4, pp. 98–102, 1996.
sifiers and imply great potential for practical applications. A [10] K. Hollingsworth, K. W. Bowyer, and P. J. Flynn, “Pupil dilation degrades
iris biometric performance,” Comput. Vis. Image Understanding, vol. 113,
pair t-test reveals no significant difference (p > 0.05) between pp. 150–157, 2009.
our method and manual classification in terms of recognition [11] J. Daugman, “The importance of being random: Statistical principles of
rates, and thus further proves the meeting of the requirements iris recognition,” Pattern Recognit., vol. 36, pp. 279–291, 2003.
[12] M. J. Hogan, J. A. Alvarado, and J. E. Weddell, Histology of the Human
for clinical applications. Eye. Philadelphia, PA: Saunders, 1971, pp. 11–14.
In further comparing the performance of the system with our [13] B. Jensen, “L’iridologie, science et pratique de l’art de guerir,” pp. 11–13,
former research, it can be seen that for a single internal organ, 1988.
[14] R. Cong, S. Song, and Y. Liu, “Study of genetic mutation locus in a
the proposed pathological features may produce a high degree family with congenital aniridia,” Chin. J. Ophthalmol., vol. 42, no. 112,
of specificity with a high recognition rate, such as the case of pp. 1113–1117, 2006.
MA et al.: IRIS-BASED MEDICAL ANALYSIS BY GEOMETRIC DEFORMATION FEATURES 231
[15] R. Roizenblatt, P. Schor, F. Dante, J. Roizenblatt, and R. Belfort, “Iris Naimin Li received the Graduate degree from the
recognition as a biometric method after cataract surgery,” Biomed. Eng. Department of Medical Treatment, Traditional Chi-
Online, vol. 3, no. 1, pp. 2–8, 2004. nese Medicine (TCM), Shenyang, China, in 1962.
[16] U.S. NLM/NIH Medline Plus, Glaucoma. (2011). [Online]. Available: He is currently a Fellow Professor at the Harbin
http://www.nlm.nih.gov/medlineplus/ency/article/001620.htm Institute of Technology, Harbin, China. Since 1965,
[17] L. MA, K. Wang, and D. Zhang, “An universal texture segmentation and he has been involved in the study and exploration of
representation scheme based on ant colony optimization for iris image the application of tongue diagnosis in modern dis-
processing,” Comput. Math. Appl., vol. 57, pp. 1862–1868, Nov. 2009. ease diagnosis and treatment, in the fields of internal
[18] R. P. Kleihorst, R. L. Lagendiik, and J. Biemond, “An adaptive order- medicine, pediatrics, epidemiology, and gynecology.
statistic noise filter for gamma corrected image sequences,” IEEE Trans. During 1970 and 1972, he was a Leader of Experts
Image Process., vol. 6, no. 10, pp. 1442–1446, Jun. 1997. Group of TCM by the Chinese government in the
[19] N. M. Kwok, Q. P. Ha, D. Liu, and G. Fang, “Contrast enhancement and Middle-Europe and Albania. In 1989, he started the first tongue image labo-
intensity preservation for gray-level images using multiobjective particle ratory in China. He has authored or coauthored more than 250 papers and 14
swarm optimization,” IEEE Trans. Autom. Sci. Eng., vol. 6, no. 1, pp. 145– books around his research areas.
155, Jun. 2009.
[20] H. Farid, “Blind inverse gamma correction,” IEEE Trans. Image Process.,
vol. 10, no. 10, pp. 1428–1433, Oct. 2001.
[21] D. Brugger, M. Bogdan, and W. Rosenstiel, “Automatic cluster detection Yan Cai received the M.S. degree from Harbin Med-
in Kohonen’s SOM,” IEEE Trans. Neural Netw., vol. 19, no. 3, pp. 442– ical University, China, in 2001 and the Ph.D. degree
459, Mar. 2008. from the Tongji Medical School at Huazhong Uni-
[22] A. K. Jain, A. Ross, and S. Prabhakar, “An introduction to biometric versity of Science and Technology in 2012.
recognition,” IEEE Trans. Circuits Syst. Video Technol., vol. 14, no. 1, She is now a Professor and Chief of Maternity De-
pp. 4–20, Jan. 2004. partment at the Fourth Affiliated Hospital of Harbin
[23] V. N. Vapnik, S. E. Golowich, and A. J. Smola, “Support vector method Medical University, Harbin, China, and the Deputy
for function approximation, regression estimation, and signal processing,” Director in Perinatology Medical Association of
Adv. Neural Inf. Process. Syst., vol. 9, pp. 281–287, 1996. HeiLongjiang Province. She has authored or coau-
[24] K. R. Tan and S. C. Chen, “Adaptively weighted sub-pattern PCA for face thored more than 30 papers and 2 books around her
recognition,” Neurocomputing, vol. 64, pp. 505–511, 2005. research areas. She is involved in gynecology and ob-
stetrics clinical work for 24 years, nearly 10 years mainly involved in obstetrics
work. Her research interests include bioinformatics, perinatology, and high-risk
pregnancy.
Prof. Cai received seven provincial R&D awards and one national patent.
Lin Ma received the M.E. degree from University She has also received the Stefano Mauritania national level medal.
Paris 6, Paris, France, in 2001 and the Ph.D. de-
gree from the Harbin Institute of Technology (HIT)
Harbin, China, in 2008, both in computer science.
Wangmeng Zuo (M’09) received the Ph.D. degree
Since 2002, she has been in the Biocomput-
in computer application technology from the Harbin
ing Research Center of Computer Science and
Institute of Technology, Harbin, China, in 2007.
Technology, HIT, where she is currently an Associate
He is an Associate Professor in the School of
Professor in the Department of Computer Science.
Computer Science and Technology, Harbin Institute
Meanwhile, in 2006 and 2011 she was a Visiting
of Technology. From July to December 2004, from
Scholar in The Hong Kong Polytechnic University
November 2005 to August 2006, and from July 2007
supported by Hong Kong Croucher Funding. Her re-
to February 2008, he was a Research Assistant in the
search interests include intelligent information processing, image processing,
Department of Computing, Hong Kong Polytechnic
and cognitive science.
University. From August 2009 to February 2010, he
was a Visiting Professor in Microsoft Research Asia.
His research interests include sparse representation, biometrics, pattern recog-
nition, and computer vision.
Dr. Zuo is an Associate Editor of IET Biometrics and the reviewers of several
David Zhang (F’09) received the Graduate degree in international journals.
computer science from Peking University, Beijing,
China, and the M.Sc. degree in computer science
and the Ph.D. degree from the Harbin Institute of Kuanquan Wang (SM’07) was born in Sichuan
Technology (HIT), Harbin, China, in 1982 and 1985, Province, China, in September 1964. He received
respectively. He received the second Ph.D. degree in the B.E. and M.E. degrees in computer science from
electrical and computer engineering from the Univer- the Harbin Institute of Technology (HIT), Harbin,
sity of Waterloo, ON, Canada, in 1994. China, and the Ph.D. degree in computer science
From 1986 to 1988, he was a Postdoctoral Fellow from Chongqing University, Chongqing city, China,
at Tsinghua University, Beijing, and then, an Asso- in 1985, 1988, and 2001, respectively.
ciate Professor at the Academia Sinica, Beijing. He From 1988 to 1998, he was with the Department
is currently a Head in the Department of Computer, and a Chair Professor at of Computer Science, Southwest Normal University,
the Hong Kong Polytechnic University, Kowloon, Hong Kong, where he is Chongqing, China. Since 1998, he has been with the
the Founding Director of the Biometrics Technology Centre supported by the Biocomputing Research Center of Computer Science
Hong Kong SAR Government in 1998. He was also a Visiting Chair Professor in and Engineering, Department of HIT, where he is currently a Professor at the
Tsinghua University, and an Adjunct Professor in Peking University, Shanghui Department of Computer Science and Engineering, and an Associate Director
Jiao Tong University, HIT, and the University of Waterloo. He is the Founder and of the Biocomputing Research Center in HIT. Meanwhile, from 2000 to 2001,
Editor-in-Chief of International Journal of Image and Graphics, Book Editor of he was a Visiting Scholar in The Hong Kong Polytechnic University supported
Springer International Series on Biometrics, an Organizer of the International by Hong Kong Croucher Funding, and from 2003 to 2004 he was a Research
Conference on Biometrics Authentication, an Associate Editor of more than ten Fellow in the same university. So far, he has published more than 90 papers.
international journals, including IEEE TRANSACTION AND PATTERN RECOGNI- His research interests include biometrics, image processing, pattern recognition,
TION. He is the author of more than ten books and 200 journal papers. and biometrics-based diagnosis technology for traditional chinese medicine.
Dr. Zhang is a Croucher Senior Research Fellow, Distinguished Speaker of Dr. Wang is an Associate Editor of International Journal of Image and
the IEEE Computer Society, and a Fellow of the International Association of Graphics, a reviewer of IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBER-
Pattern Recognition. NETICS, Pattern Recognition.