Computerized Medical Imaging and Graphics 35 (2011) 179–185
Contents lists available at ScienceDirect
Computerized Medical Imaging and Graphics
journal homepage: www.elsevier.com/locate/compmedimag
Sublingual vein extraction algorithm based on hyperspectral tongue
Qingli Lia,∗, Yiting Wangb, Hongying Liua, Yana Guana, Liang Xua
aKey Laboratory of Polor Materials and Devices, East China Normal University, Shanghai, China
bAdvanced Institute of Drug Discovery and Development, East China Normal University. Shanghai, China
a r t i c l ei n f o
Received 22 December 2009
Received in revised form
22 September 2010
Accepted 1 October 2010
Hidden Markov model
a b s t r a c t
may have pathological relationship with some diseases. To analyze this information quantitatively, one
primitive work is to extract sublingual veins accurately from tongue body. In this paper, a hyperspectral
tongue imaging system instead of a digital camera is used to capture sublingual images. A hidden Markov
model approach is presented to extract the sublingual veins from the hyperspectral sublingual images.
This approach characterizes the spectral correlation and the band-to-band variability using a hidden
and the spectral angle mapper algorithm are tested on a total of 150 scenes of hyperspectral sublingual
the proposed algorithm can extract the sublingual veins more accurately than the traditional algorithms
and can perform well even in a noisy environment.
Crown Copyright © 2010 Published by Elsevier Ltd. All rights reserved.
The human tongue is one of the most important organs of the
body that carries an abundance of information about the human
health status . Whenever there is a complex disorder full of
contradictions, an examination of the tongue can instantly clarify
the main pathological process. Therefore, the examination of the
tongue is one of the most important approaches in obtaining sig-
nificant evidence when diagnosing a patient’s health condition. So
far, most inspections of the tongue have been focused on shape,
color, texture, coating, and cracks [2–4]. Aside from these various
tongue. Sublingual veins refer to the veins distributed on the lower
surface of the tongue directly connected with the viscera organs
and the blood through certain channels . The inspection of the
sublingual veins can provide valuable insights into the health sta-
tus of the human body. However, the subjective characteristics of
the traditional method impedes this objective because sublingual
vein diagnoses are usually based on detailed visual discrimination,
which mainly depends on the subjective analysis of the examiners
∗Corresponding author at: Key Laboratory of Polor Materials and Devices, East
China Normal University, School of Information Science, No. 500, Dongchuan Rd.,
Shanghai, China. Tel.: +86 0 21 5434 5199; fax: +86 0 21 5434 5199.
E-mail address: firstname.lastname@example.org (Q. Li).
Nowadays, the rapid progress of information technology pro-
motes the automatization of tongue disease diagnosis based on
modern image processing and pattern recognition approaches
[7,8]. For example, Li and Yuen  investigated the color matching
of tongue images in different color spaces with different metrics.
Zhang et al. presented methods for extracting tongue cracks ,
automated tongue body segmentation , tongue shape classi-
fication , and computerized tongue diagnosis . Moreover,
Chiu  built a computerized tongue examination system based
on computerized image analysis to quantize the tongue proper-
ties in traditional Chinese medical diagnosis. These studies prove
that the accurate extraction of tongue features is important for
computerized tongue diagnosis. The automatic extraction of sub-
lingual veins from complex scenes should also be foremost solved
due to the qualities of segmentation directly influencing the subse-
quent feature extraction and recognition. Some experiments have
era under a visible light source or an infrared light source for the
extraction of sublingual veins. Takeichi and Sato  performed
computer-assisted image analyses on the color of the tongues of
95 medical students to enhance the accuracy and objectivity of
tongue inspections for determining blood stasis. They selected
those areas containing sublingual veins from the tongue body
image and subjected them to image analysis. A slide scanner was
used to digitize the color slides, and the digital information was
transmitted to a personal computer for subsequent feature extrac-
tion and analysis. This is a prior research on color of sublingual
0895-6111/$ – see front matter. Crown Copyright © 2010 Published by Elsevier Ltd. All rights reserved.
Q. Li et al. / Computerized Medical Imaging and Graphics 35 (2011) 179–185
 Baum LE, Petrie T, Soules G, Weiss N. A maximization technique occurring in
the statistical analysis of probabilistic functions of Markov chains. Ann Math
 Baum LE. An inequalityand associated maximization technique in statisti-
cal estimation for probabilistic functions of Markov processes. Inequalities
 Rabiner LR. A tutorial on hidden Markov models and selected applications in
speech recognition. Proc IEEE 1989;72(2):257–86.
 Rosti A-VI, Gales MJF. Factor analysed hidden Markov models for speech recog-
nition. Comput Speech Lang 2004;18(2):181–200.
 Park H-S, Lee S-W. A truly 2-D hidden Markov model for offline handwritten
character recognition. Pattern Recognit 1998;31(12):1849–64.
 Bunke H, Roth M, Schukat-Talamazzini EG. Off-line cursive handwriting recog-
nition using hidden markov models. Pattern Recognit 1995;28(9):1399–413.
 Chen F-S, Fu C-M, Huang C-L. Hand gesture recognition using a real-time track-
 Keshet J, Shalev-Shwartz S, Singer Y, Chazan D. A large margin algorithm for
speech-to-phoneme and music-to-score alignment. IEEE Trans Audio Speech
Lang Process 2007;15(8):2373–82.
 Grindlay G, Helmbold D. Modeling, analyzing, and synthesizing expressive
piano performance with graphical models. Mach Learn 2006;65(2–3):361–87.
 Llorens C, Futami R, Renaud G, Moya A. Bioinformatic flowchart and database
to investigate the origins and diversity of Clan AA peptidases. Biol Direct
 Martin J, Gibrat J-F, Rodolphe F. Analysis of an optimal hidden Markov model
for secondary structure prediction. BMC Struct Biol 2006;6(25).
 Tai AH, Ching W-K, Chan LY. Detection of machine failure: Hidden Markov
Model approach. Comput Ind Eng 2009;57:608–19.
 Chan LY, Lai CD, Xie M, Goh TN. A two-stage decision procedure for monitoring
processes with low nonconforming fraction. Eur J Oper Res 2003;150:420–36.
 Ledauphin S, Pommeret D, Qannari EM. Application of hidden Markov model
to products shelf lives. Food Qual Prefer 2008;19(2):156–61.
 Gong X, Huang N-K. Textured image recognition using hidden Markov model
ICASSP 88. In: 1988 International conference on acoustics, speech, and signal
processing. New York, NY, USA: IEEE; 1988. p. 1128–31.
 He Y, Kundu A. Shape classification using hidden Markov model ICASSP 91. In:
1991 International conference on acoustics, speech and signal processing. New
York, NY, USA: IEEE; 1991. p. 2373–6.
 Satish L, Gururaj BI. Use of hidden Markov models for partial discharge pattern
classification. IEEE Trans Electr Insul 1993;28(2):172–82.
 Poggi G, Ragozini A. Image segmentation by tree-structured Markov random
fields. IEEE Signal Process Lett 1999;6(7):155–7.
 Choi H, Baraniuk R. Multiscale image segmentation using wavelet-domain hid-
den Markov models. IEEE Trans Image Process 2001;10(9):1309–21.
 Kato J, Watanabe T, Joga S, Rittscher J, Blake A. An HMM-based segmenta-
tion method for traffic monitoring movies. IEEE Trans Pattern Anal Mach Intell
 Lim J, Pyun K. Cost-effective hidden Markov model-based image segmentation.
IEEE Trans Process Lett 2009;16(1–3):172–5.
 Yu SZ, Kobayashi H. An efficient forward-backward algorithm for an explicit-
duration hidden Markov model. IEEE Signal Process Lett 2003;10(1):11–4.
 Kruse FA, Lefkoff AB, Boardman JW, Heidebrecht KB, Shapiro AT, Barloon PJ, et
al. The spectral image processing system (SIPS)—interactive visualization and
analysis of imaging spectrometer data. Remote Sens Environ 1993;44:145–
 Kruse FA, Lefkoff AB, Boardman JW. The spectral image processing system
(SIPS)-software for integrated analysis of AVIRIS data. In: Sum. 4th Ann. JPL
Airborne Geosci. Workshop. Pasadena: JPL Pub.; 1992. p. 23–5.
 Li QL, Liu Z. Tongue color analysis and discrimination based on hyperspectral
images. Comput Med Imaging Graphics 2009;33(3):217–21.
 Ballard DH, Brown CM. Computer vision. Prentice Hall; 1982.
 Liu LL, Zhang D, Kumar A, Wu X. Tongue line extraction. Pattern Recognition,
2008. 19th International conference on ICPR 2008. Tampa, FL, 2008. p. 1–4.
 Pham BL, Cai Y. Visualization techniques for tongue analysis in traditional Chi-
nese medicine. In: Galloway RL, editor. Proc SPIE, 2004. p. 171–80.
the 3rd international symposium on image and signal processing and analysis,
ISPA 2003, 2003. p. 626–29.
Qingli Li received the B.S. and M.S. degrees in computer science and engineering
from Shandong University, Jinan, China, in 2000 and 2003, respectively, and the
Ph.D. degree in pattern recognition and intelligent system from Shanghai Jiaotong
Materials and Devices, East China Normal University, Shanghai, China, where he is
Yiting Wang received the B.S. degree from Fudan University, Pharmacology, Shang-
hai, China, in 1984, and the Ph.D. degree from Hamamatsu University School
Medicine, Hamamatsu, Japan, in 1992. From 1984 to 1986, she was an Apothecary
of Shanghai Institute for Drug and Food Control. From 1992 to 1998, she was a
Researcher of AkaneKanbou Pharmacy, Inc., Japan. From 1994 to 1995, she was a
Visitor of Edinburgh University. From 1998 to 2005, she was the Deputy Director of
Shanghai Institute for Drug and Food Control. Since 2005, she has been a Professor
at the Advanced Institute of Drug Discovery and Development, East China Normal
University, Shanghai, China. She is the author or co-author of more than 50 papers
research interests include bio-imaging, drug delivery, biosensors, DNA, and protein
Hongying Liu received the B.S. degrees in automatic control from Nanjing Univer-
sity of Aeronautics and Astronautics, and M.S. degrees in circuits and systems from
South East University, nanjing, China, in 1997 and 2003, respectively. Now she is
doing her best to gain Ph.D. degree in communication engineering from East China
Materials and Devices, East China Normal University, Shanghai, China, where she is
Yana Guan received the B.S. degree in the school of physical science and tech-
nology from Anhui University, Hefei, China, in 2009. She is currently pursuing the
M.S. degree at the Key Laboratory of Polor Materials and Devices, East China Nor-
analysis, pattern recognition.
Liang Xu is currently pursuing the M.S. degree at the Key Laboratory of Polor
Materials and Devices, East China Normal University, Shanghai, China. His research
interests include hyperspectral image analysis, pattern recognition.