SS symmetry

Article
Classification of Body Constitution Based on TCM
Philosophy and Deep Learning
Yung-Hui Li , Muhammad Saqlain Aslam *, Kai-Lin Yang, Chung-An Kao and Shin-You Teng
Department of Computer Science and Information Engineering, National Central University, Taoyuan 32001,
Taiwan; yunghui@csie.ncu.edu.tw (Y.-H.L.); 106522001@cc.ncu.edu.tw (K.-L.Y.);
105502561@cc.ncu.edu.tw (C.-A.K.); 105502011@cc.ncu.edu.tw (S.-Y.T.)
* Correspondence: saqlain@g.ncu.edu.tw; Tel.: +886-3-4227151-35242




Received: 8 April 2020; Accepted: 6 May 2020; Published: 12 May 2020


Abstract: There is a growing demand for alternative or complementary medicine in health care
disciplines that uses a non-invasive instrument to evaluate the health status of various organs inside
the human body. In this regard, we proposed a real-time, non-invasive, and painless technique to
assess an individual’s health condition. Our approach is based on the combination of iridology and
the philosophy of traditional Chinese medicine (TCM). The iridology chart presents perfect symmetry
between the left and right eyes, and such a unique representation reveals the body constitution based
on TCM philosophy, which classifies the aforementioned body constitution into a combination of
nine categories to describe the varieties of genomic traits. In addition, we applied a deep-learning
method along with the combination of iridology and TCM to predict the possible physiological
or psychological strength or weakness of the subjects and give advice to them about how to take
care of their health according to the body constitution assessment. We used several pre-trained
convolutional neural networks (CNNs, or ConvNet), such as a residual neural network (ResNet50),
InceptionV3, and dense convolutional network (DenseNet201), to classify the body constitution using
iris images. In the experiments, the CASIA-Iris-Thousand database was used to perform this task.
The experimental results showed that the proposed iris-based health assessment method achieved an
82.9% accuracy.

Keywords: iridology; health care; deep learning; complementary medicine; TCM; five elements

1. Introduction
In recent years, there has been an increasing interest in non-invasive, real-time, and painless
approaches for health status assessment. Because of its painless and non-invasive nature, it has gained
tremendous attention and demand in the various healthcare fields.
The human eye plays a pivotal role in such approaches due to having a unique appearance, shape,
and color of the iris. The iris is considered an internal human body organ, but externally, it may be easily
observed [1]. Each person has a different spatial shape of the iris; therefore, the iris texture has been
used extensively for biometrics for the last several decades [1–3]. Besides the biometric identification,
iris textures and colors are related to the healthiness of the internal organs of a person. The place
of the iris is near the human nervous system, and after examining the iris, it can reveal significant
information about the human health imbalance and nature of possible disturbances in internal body
organs, which is called iris diagnosis (iridology) [2,4]. Additionally, extensive research has shown
that the human iris can be used as a manifestation of health status [2,3]. Therefore, the human eyes
are referred to as the window to the soul and everything related to health is encrypted in our eyes.
The iris has definite characteristics, e.g., fine lines, crypts, and lesions, which are associated with the
body constitution, inherent weaknesses, health levels, and changes that occur in the human body [5].

Symmetry 2020, 12, 803; doi:10.3390/sym12050803 www.mdpi.com/journal/symmetry

Symmetry 2020, 12, 803 2 of 17
Symmetry 2020, 12, x FOR PEER REVIEW 2 of 17

The irishuman
in the is formed before
body [5]. birth andisit formed
The iris remains before
the same throughout
birth one’s lifetime
and it remains [6].throughout
the same The iris is aone’s
thin
spherical plate between the pupil and cornea, as shown in Figure 1.
lifetime [6]. The iris is a thin spherical plate between the pupil and cornea, as shown in Figure 1.

Figure1.1. Human
Figure Human iris
iris image
image[7].
[7].

Iridology,
Iridology, also also knownknown as as iridodiagnosis
iridodiagnosis or or iridiagnosis,
iridiagnosis, is is an
an approach
approach that that uses
uses thethe patterns,
patterns,
colors,
colors, and many manyother othervisible
visiblecharacteristics
characteristics of the
of the iris iris to assess
to assess an individual’s
an individual’s general general
healthhealth
status.
status.
Such visible characteristics include pigmentary change, color, and pathological lesions, etc. [8].[8].
Such visible characteristics include pigmentary change, color, and pathological lesions, etc. In
In iridology,the
iridology, theirisirisstructure,
structure,color,
color,and
andunique
uniquemarkings
markings provide provide useful
useful information about about thethe health
health
status
status [4].
[4]. Iridologists
Iridologists believebelieve that
that every
every organ
organ in in the
the human
human body body has has its
its corresponding
corresponding area area ininthe
the
iris. Therefore, the assumption is that the status of body organs can
iris. Therefore, the assumption is that the status of body organs can be observed by simply examining be observed by simply examining
the
thecorresponding
correspondingregions regionsof ofthe
theiris.
iris.
According
Accordingto toiridologists’
iridologists’investigations,
investigations,when whenany any change
change happens
happens in in the
the human
human body, body,such
suchas as
weaknesses and strengths, a spot or mark appears in the iris. These marks
weaknesses and strengths, a spot or mark appears in the iris. These marks on the iris indicate possible on the iris indicate possible
diseases
diseasesor ordysfunctions
dysfunctionsofofspecific specific body
bodyorgans
organs [5].[5].
ForForinstance,
instance,changes
changes in kidneys
in kidneys willwill
appear at the
appear at
bottom edge of the iris. The right kidney affects the right iris, while the
the bottom edge of the iris. The right kidney affects the right iris, while the left kidney affects the left left kidney affects the left iris.
Changes
iris. Changesin theinstomach
the stomach can cause changes
can cause in theiniris
changes thejust
irisabove the pupil.
just above Thus,Thus,
the pupil. iridology reveals
iridology any
reveals
transition
any transition in a human
in a human body organ
body because the iris the
organ because has iris
many hasregions
many that are associated
regions with different
that are associated with
body
differentparts. Theparts.
body spot or Themark
spotwill not affect
or mark the affect
will not result the of the iris of
result recognition because there
the iris recognition is a fault
because there
tolerance
is a faultmechanism
tolerance built in the iris
mechanism recognition
built in the irisalgorithm.
recognition For the iris recognition
algorithm. For the algorithm, most of
iris recognition
the time, researchers
algorithm, most of the settime,
the hamming
researchers distance
set the(HD) threshold
hamming value(HD)
distance to 0.33 for the matching
threshold criteria
value to 0.33 for
(HD =
the matching criteria (HD = 0 means perfect match) and this will tolerate a little bit of mismatch is
0 means perfect match) and this will tolerate a little bit of mismatch in the iris texture. This in
why iridologists
the iris texture. This claimisthatwhythe iris color may
iridologists claimchange
that the but it color
iris will not may affect iris recognition.
change but it will not affect iris
In order to identify the corresponding region for different body organs, two schematic diagrams
recognition.
have beenIn orderdeveloped.
to identify First,
thethere are seven concentric
corresponding region for zonesdifferent in the
body iris, and the
organs, twopupil is their diagrams
schematic common
axis,
haveas shown
been in Figure
developed. 2 [9].
First, Theare
there ring nearest
seven to the zones
concentric pupil reflects
in the iris, theand
healthiness
the pupilcondition of the
is their common
stomach,
axis, as shown in Figure 2 [9]. The ring nearest to the pupil reflects the healthiness condition zones
whereas the most external circle reflects the skin. Second, the iris is divided into radical of the
(also known
stomach, as an iris
whereas thechart).
most There
externalarecircle
12 equal sections
reflects the radiating
skin. Second, out from the pupil
the iris to theinto
is divided iris border,
radical
which
zones are (alsosimilar
known toas the
anface
iris of a clock,
chart). as shown
There in Figure
are 12 equal 3 [9].radiating out from the pupil to the iris
sections
The iris chart in Figure 3 is very helpful
border, which are similar to the face of a clock, as shown in Figure in identifying imbalanced3 [9]. body parts. The iris chart
divides the iris surface into a number of zones, and each segment is related to an inner organ or a
human body function [4,10,11]. From this chart, we can see the corresponding regions of the organs,
such as the digestive system, brain, kidney, and heart. Amazingly, the iridology chart for the left and
right eyes is symmetric, demonstrating the mysterious balancing nature of the body.
Conventionally, iridologists use a camera and microscope to examine the subject’s iris and then
observe any irregularities based on their professional experience. Finally, practitioners compare their
observations to the iris chart.
Symmetry 2020, 12, x FOR PEER REVIEW 3 of 17

Symmetry 2020, 12, 803 3 of 17

Symmetry 2020, 12, x FOR PEER REVIEW 3 of 17

Figure 2.2.
Figure
Figure 2.The
Theseven
The seven zonesof
seven zones
zones ofof
thethe
the iris.iris.
iris.

Figure 3. Iridology chart is shown for both the right and left iris and their relations to body organs
[5].

Figure Figure
3. Iridology
The iris chartchart is shown
3. Iridology chart
in Figure
is shownfor
3 is very both theright
for both the
helpful right
in and
and left left
identifying iristheir
iris and andrelations
imbalanced their relations
body to body
to body organs
parts. The
[5]. organs
iris chart
divides the iris surface into a number of zones, [5] and. each segment is related to an inner organ or a
Furthermore, iridology has become particularly the most valuable and widely used diagnostic
human
tool body function [4,10,11].
in complementary From this
and alternative chart, (CAM),
medicine we can see the corresponding
especially in the Unitedregions of the
States and organs,
Germany,
Thewhere
iris chart
such as the
there arein several
Figureiridology
digestive system,3 is very
brain, helpful
kidney,
institutes andin
and identifying
heart.
societiesAmazingly, imbalanced medicine
the
[12]. Complementary iridology body parts.
chart for the The
left
is a research andiris chart
rightiris
divides the
branch eyes is symmetric,
surface
consisting into
of demonstrating
a number
diverse medicalof the
and mysterious
zones,
health and balancing
each
care systems thatnature
segment claimisof the body.tothe
torelated
improve anstandard
inner organ of or a
health,Conventionally,
prevent disease, iridologists
and use
address a camera
conditions and microscope
that orthodox to examine
medicine the
has subject’s
had limited irissuccess
and then in
human body function [4,10,11]. From this chart, we can see the corresponding regions of the organs,
observe anylike
remedying, irregularities
chronic back based painonand
their professional
certain cancers experience.
[13]. Today,Finally,
in the practitioners
USA and Europe, comparethefor their
use of left and
such as the digestive
observations to the system, brain,
iris chart. kidney, and heart. Amazingly, the iridology chart the
complementary medicine in health care has increased; for instance, around two-thirds of the general
right eyes is Furthermore,
symmetric, iridologydemonstrating hasSome
become theparticularly
mysterious thebalancing
most valuable nature of the used
and widely body. diagnostic
population now use CAM [14]. therapists practicee iridology as a basis for prescribing dietary
Conventionally,
tool in complementary
supplements [15]. iridologists
and use a
alternative camera
medicine and microscope
(CAM), especially to
in examine
the United the
States subject’s
and Germany, iris and then
where there
observe any Iridology are
irregularities several iridology
based
also plays institutes
on their
a vital and
professional
role in societies [12].
experience.
traditional Chinese Complementary
medicine Finally,
(TCM) medicine
practitioners is
treatment. TCMa research
compare is their
branch
a type consisting
of CAM of diverse
that practices medical and health
customized care systems
therapies that arethat
basedclaim on tobody
improve the standard
constitution theory.of
observations to the iris chart.
health, prevent
However, TCM disease,
theory isand based address conditions that
on a philosophical orthodox medicine
framework, has had limited success in
Furthermore,
remedying, iridology
like chronic has
back become
pain and particularly
certain cancers [13]. moste.g.,
the Today, in
the five-element
valuable
the USA and
and widelypostulate
Europe, used
the use
[16].
diagnostic
of
Ancient Chinese philosophers hypothesized that nature is composed of five primary elements, similar to
tool in complementary
complementary and alternative
medicine in health medicine (CAM), forespecially in thetwo-thirds
United States and Germany,
how Greek scholars presented the care has increased;
creation of all the elementsinstance, around
in the universe. TCMof the general
classifies the
where there
human areorganic
population several
nowsystem iridology
use CAM into[14]. institutes
five Some
organs and
therapists
and societies
links practicee
them with [12].
fiveComplementary
iridology as a basis
elements. Theforfive medicine
prescribing
elements oris
dietary a research
five
supplements
branch consisting
states are sorted [15].
of diverse
in ordermedical
according and health
to five care systems
materials, such as wood, that claim to improve
fire, earth, metal, and the standard of
water.
Iridology
Every component also plays a vital role in traditional Chinese medicine (TCM) treatment. TCM is a type
health, prevent disease,has and itsaddress
own literal and logicalthat
conditions meaning [17]. TCM
orthodox practitioners
medicine has had observe human
limited success in
of CAM
eyes as a that practices
diagnostic customized
tool. Physicians therapies
who that are
practice TCMbasedhave onput
body the constitution
five-components theory. However,
assumption
remedying, like
TCMpractice, chronic
theory is basedback onit pain and certain
a philosophical cancers [13].the
framework, Today, in the postulate
USA and[16]. Europe, the use of
into spreading to clinical practice as a kind ofe.g., five-element
five-element acupuncture (FEA), which Ancientwas
complementary medicine in
Chinese philosophers health care has increased; for instance, around two-thirds of the general
presented by a British hypothesized
acupuncturist,that Dr. nature is composed
J.R. Worsley [17,18].of five primary elements, similar to how
populationGreek now use CAM
scholars
Generally, presented
more [14].
than theSome
one therapists
creation of allof
characteristic thethepracticee
elements iniridology
five elements thecan
universe. as TCM
be observeda basis forsame
classifies
in the prescribing
human dietary
theperson’s
organic
supplements system into five organs and links them with five elements. The
[15].Based on the suggestions from the experts of Taiwan International Institute of Iridology
iris image. five elements or five states
are sorted in order according to five materials, such as wood, fire, earth, metal, and water. Every
Iridology also plays a vital role in traditional Chinese medicine (TCM) treatment. TCM is a type
component has its own literal and logical meaning [17]. TCM practitioners observe human eyes as a
of CAM that practices customized therapies that are based on body constitution theory. However,
TCM theory is based on a philosophical framework, e.g., the five-element postulate [16]. Ancient
Chinese philosophers hypothesized that nature is composed of five primary elements, similar to how
Greek scholars presented the creation of all the elements in the universe. TCM classifies the human
Symmetry 2020, 12, 803 4 of 17

(TIII), for each person, the body constitution can be classified into a primary type and secondary type
of the five materials. Therefore, the five elements of categorization are not exclusive. They can overlap
and co-exist in one iris. Therefore, we adopted the categorization suggested by TIII to derive nine
categories of body constitution classification for each subject [19]. These nine categories are formed by
combining five elements in an order, such as primary and secondary types, as shown in Table 1.

Table 1. Categorization of nine subclasses of body constitution based on Traditional Chinese Medicine
(TCM) philosophy.

Primary Secondary
Metal
Earth Wood
Water
Metal
Fire Wood
Water
Water Wood
Wood Metal
Metal Water

In this study, we introduce a non-invasive iris-based health assessment system that adopts the
views of experts from TIII to classify the human body into nine classes, which may give health care
advice to a subject based on body constitution evaluation. We apply deep learning techniques that
can categorize a subject and possible organ dysfunction based on the iris images. Deep learning (DL)
is so far the most powerful and robust classification technique among all of the machine learning
methods. DL has received significant interest due to its supremacy regarding accuracy when trained
with a large volume of data. Starting from 2012 to 2017, the campion of the ImageNet competition
was always won by DL approaches, especially the convolutional neural network (CNN) [20–24].
Furthermore, we utilize several existing CNN models to classify body composition using the iris image.
In recent years, CNN has made the most significant novelties in the field of computer vision and
image processing and gained the researcher’s attention. A CNN is a particular kind of multi-layer
neural network, built to identify visual patterns directly from raw images with the least preprocessing.
By using DL and CNN, we can achieve end-to-end learning, which means we do not need to care too
much about image pre-processing. As long as we collect enough data and send it to the network, it will
learn by itself. In this way, it saves a lot of time for the researcher.
The proposed system takes a colored iris image of the subject, processes the image using the
proposed DL algorithm, classifys the subject’s body constitution into one of the nine categories,
and brings up the warnings about the potential organ disorder with suggestions and advice.
In our experiments, the CASIA-Iris-Thousand database was used to perform training, testing,
and cross-validation for the proposed CNN models.
The motivation of this work was to develop an online useable practical system that can be used in
clinics and homes to perform body assessment without the need to visit a doctor or a hospital and
get results immediately. With our system, medical and travel expenses can be greatly reduced and
assessment of the health condition can be sped up.
The major contribution of this study is that we combined the image processing technique based on
the iris recognition framework (including iris image acquisition, image processing, iris segmentation)
and the powerful image classification technique based on CNN (including ResNet 50, Inception V3,
and DenseNet 201) to create a practically useful system, which is able to capture iris images and
perform nine body constitution classifications in real-time, with a classification accuracy up to 82.9%.
Such a system is interdisciplinary research consisting of iris image processing, DL, TCM, and iridology.
This paper makes the following contributions:
Symmetry 2020, 12, 803 5 of 17

# We built the first largest scale iridology database using CASIA-Iris-Thousand, which contains
20,000 iris images, with help from TIII experts.
# The powerful DL model was applied to perform body constitution classification and was able to
achieve a very high accuracy.
# A practically proven system was built, which can work in a real-time environment. This shows
that the designed system is not only theoretically superior but practically useful.
The paper is organized into the following sections. The related work is described in Sections 2
and 3 discusses the materials and methods; Section 4 presents the experimental results; Section 5 is
about the demonstration of the system; and Section 6 draws conclusions.

2. Related Work
Several studies have been performed on human eye iris examinations (iridology) to evaluate the
human body organ abnormalities and diagnostic effectiveness of iridology.
Othman et al. [4] presented preliminary research on iris recognition methods to identify
abnormalities in the human body organs. They used a novel iris segmentation approach termed the
water flow method due to its geometrical resilience and topological rubberiness. Though, the study
was preliminary research work with no significant conclusion. Hussain et al. [5] proposed an iris
algorithm for obstructive lung disease (OLD) by a non-invasive diagnosing approach. They developed
a real-time iris-based lung pre-diagnostic system, which uses a machine learning algorithm and an
iridology chart to detect lung disorder. Features were obtained by using a Gabor filter and support
vector machine (SVM). In this study, 50 lung patients and 50 healthy subjects were examined, and the
system accuracy for lung disorder detection was 88%.
Anna et al. [25] introduced an adaptive resonance theory 1 (ART1), a type of artificial neural
network that uses unsupervised learning algorithms to analyze lung disorder through an iris image.
In this paper, they designed a lung imbalance detection system that contains several phases, such as iris
segmentation, extraction of color dissimilarities, a transformation of the lung and pleura representation
region in the iris image as the input of ART1, and pattern recognition by the ART1 neural network
framework. In the changing environment of pattern recognition, ART1 is considered one of the most
durable and flexible solutions. Adelina et al. [26] proposed a method to identify diabetes by analyzing
pancreatic signs through iridology. A backpropagation method was used to identify the condition of the
pancreas organ. A Gaussian filter was used to minimize the noise on the image while iris segmentation
was done by using the Hough circle transfer method. They used the gray level co-occurrence matrix
(GLCM) for region of interest (ROI) feature extraction.
Hussain et al. [13] studied the diagnostic effectiveness of iridology to diagnose abnormalities
of the kidneys. They designed an automated approach with an artificial intelligence technique
that goes through several steps starting from iris image acquisition, pre-processing, normalization,
segmentation, feature extraction, and an adaptive neuro-fuzzy inference system (ANFIS) method for
the classification of features from a normal and chronic renal failure subjects group. They examined
data from 168 individuals free from kidney disease and 172 patients with chronic renal failure.
They obtained an 82% and 93% classification accuracy for normal and kidney patients, respectively.
The proposed method was performed on a systemic disease with ocular manifestations that showed
promising results. Lim et al. [8] conducted a longitudinal study to investigate the validity of iris
parameters, temperament characteristics (TCs), and the association among them over 1 year. They use
an intra-class coefficient (ICC) to examine inter-rater consistencies and the robustness of the iris
parameters. They performed decision tree modeling that used the iris parameters as a component for
high-level and low-level temperament features. This research has some drawbacks, such as it was
accomplished only in healthy subjects, and the duration of 1 year could not ensure the efficacy of the iris
ever. Hernandez et al. [6] investigated alternative techniques of Alzheimer’s disease detection from the
iris image using digital image processing. They used a mathematical model (based on Matlab software)
to learn about the presence of Alzheimer’s disease by analyzing the iris pattern. In their proposed
Symmetry 2020, 12, 803 6 of 17

model, a Fourier transform was used to normalize the iris image and Hough transform for the detection
of the pupil and iris in an image. Several methods based on three multilayer classifiers were used,
including naive Bayes, ZeroR, and multilayer perceptron. The performance of the naive Bayes classifier
was better than others and achieved a 61.96% diagnostic accuracy. Permatasari et al. [27] proposed a
method of heart condition detection from iris classification using the support vector machine (SVM).
Feature extraction was done by using a principal component analysis (PCA) method. They obtained a
classification accuracy rate of 80%. Herlambang et al. [28] introduced liver disorder in the iris image.
They proposed a liver disease detection technique using a back-propagation neural network and gray
level co-occurrence matrix (GLCM) for feature extraction. They used Matlab to build a liver disorder
diagnosis application. In this study, 60 individuals’ images of the right eye iris were taken, of which 34
were healthy iris images and 26 were abnormal images.
The validity of this approach was tested on 35 liver patients and obtained a 91.42% accuracy rate
in detecting liver disease. Miranda et al. [29] developed an automated algorithm technique to detect
intestinal system pathologies, such as a distended colon, healthy intestinal tract, prolapse of transverse
colon, and spastic colon, by using the human iris parameters. However, this was a pilot study, and thus
requires further examination of its efficacy.
In this paper, we evaluated the human organic system based on the concept of TIII experts and
iridology together. Furthermore, we applied powerful DL classification algorithms to iridology and
built a practically useful system for the general public. All the previous methods mainly focused on
specific body organ dysfunction detection and used a smaller amount of iris images in the experiments.
On the other hand, our approach performed a holistic health assessment by using the largest scale
dataset of iris images. Moreover, the designed system is a real-time, user-friendly, and portable device
that can perform the assessment very quickly. The summary of the proposed method and the existing
studies of iridology is shown in Table 2. Since the classification goal and the datasets used in our work
are different than any of the prior studies, it is inappropriate to directly compare our work with the
prior works. Thus, Table 2 is simply a summary that lists all related prior works and also our work.

Table 2. Summary of the existing studies using iridology and the proposed method.

Number of Number of Disease Accuracy

Methods
Classes Subjects Type %
Feature extracted by Gabor filter and SVM Hussain et al. [5] 2 100 Lung 88
Adaptive Resonance Theory (ART) Anna et al. [25] 2 32 Lung 91.40
Feature extraction by Gray Level Co-Occurrence Matrix
2 20 Diabetes 81.35
(GLCM) Adelina et al. [26]
Adaptive Neuro-Fuzzy Inference System (ANFIS) for
2 340 Kidney 82/93
classification Hussain et al. [13]
Decision tree model for feature extraction Lim et al. [8] 2 70 - 77.5/88.9
Fourier Transform for normalization, and Hough Transform
for detection of a pupil. For classification: Naive Bayes, 2 - Alzheimer’s 61.96
ZeroR, and multilayer perceptron Hernandez et al. [6]
Feature extracted by (PCA) method and Support Vector
2 40 Heart 80
Machine (SVM) for Classification Permatasari et al. [27]
Feature extraction by Gray Level Co-Occurrence Matrix
2 35 Liver 91.42
(GLCM) Herlambang et al. [28]
Deep learning edge-based algorithms for Iris Holistic
segmentation. CNN models for Classification 9 1000 Health 82.9
(Proposed Method) Assessment
Symmetry 2020, 12, 803 7 of 17

3. Materials and Methods

This section describes the proposed real-time iris-based health assessment system according to
the view of TCM and iridology.

3.1. Iridology with Traditional Chinese Medicine (TCM)

Traditional Chinese medicine (TCM) is a conventional style of Chinese medicine practice, which is
based on a unique hypothesis along with diagnostic and treatment approaches [30] and has a history of
more than 2500 years [31]. This is the principle of Chinese philosophy, history, and medical knowledge.
According to the speculations of ancient Chinese scholars, the creation of the world rotates around the
five fundamental elements of daily life, and the formation of the human body is the combination of
the five elements. Each element represents a dissimilar attribute, function, or appearance, according
to which everything in the universe can be classified. Therefore, TCM practitioners categorize the
inner body functioning system into a five-organ system, namely the heart, liver, kidney, stomach
(including spleen), and lung, which are further linked to fire, wood, water, earth, and metal, respectively,
based on the morphological or functional similarity [17]. Table 3 presents the correspondence between
the five elements and the human body organs.

Table 3. Five elements of Traditional Chinese medicine (TCM) and its correspondence to the human
internal organs.

Elements of TCM Corresponding Human Organs

Wood Liver
Water Kinder and Bladder
Metal Lungs and Large Intestine
Earth Stomach and Spleen
Fire Heart and Small Intestine

In TCM, the concept of five elements has been practiced to describe the universe and the
connections of the main organs of the human body [32]. The imbalance between the five elements is
supposed to cause diseases.
In this study, we propose a practical health assessment system that combines DL with the
philosophy of TCM and iridology, which evolves into a convenient and portable tool that is able to
predict possible organ dysfunction and give advice to subjects about how to take care of their health
according to the body constitution assessment. According to the advice of iridology experts from TIII,
the key to body constitution using iridology lies in the third zone (ring) of the iris chart. We summarized
the classification rules and present them in the form of a table. Table 4 shows the classification of five
elements based on the third zone (ring) of the iris chart. Figure 4 shows 15 examples of iris images
corresponding to five TCM elements, based on the classification criterion in Table 4. Figure 5 shows the
samples of iris images of the derived nine categories (with primary and secondary type), which were
used in our experiments.

Table 4. Classification criteria of the five body constitutions based on the third zone (ring) of the iris
chart (autonomic nervous system).

TCM Elements Ring Shape

Wood It has no explicit ring pattern.
Water It has an explicit ring pattern and very small acute angles acceptable.
Metal It has acute angles around itself.
Earth Its radius is bigger than the radius of 1/3 iris.
Fire Its radius is smaller than the radius of 1/3 iris.
Symmetry 2020,12,
Symmetry2020,
Symmetry 2020, 12,803
12, xxFOR
FORPEER
PEERREVIEW
REVIEW 888 of
of
of17
17
17

(a)
(a) (b)
(b) (c)
(c) (d)
(d) (e)
(e)
Figure
Figure 4.4.Samples
Figure4. of
Samplesof
Samples ofiris
irisimages
iris based
imagesbased
images on
basedon the
onthe third
thethird ring
thirdring on
ringon the
onthe iris.
iris.(a)
theiris. (a)Wood
(a) type
Woodtype
Wood iris
typeiris image.
image.(b)
irisimage. (b)Water
(b) Water
Water
type iris
typeiris
type image.
irisimage. (c)
image.(c) Metal
(c)Metal type
Metaltype iris
typeiris image.
image.(d)
irisimage. (d)Earth
(d) type
Earthtype
Earth iris
typeiris image.
image.(e)
irisimage. (e)Fire
(e) Firetype
Fire typeiris
type irisimage.
iris image.
image.

(a)
(a) (b)
(b) (c)
(c) (d)
(d) (e)
(e) (f)
(f) (g)
(g) (h)
(h) (i)
(i)
Figure
Figure5.5.Samples
Figure Samplesof
Samples ofiris
irisimages
imagesof
images ofthe
of thederived
the derivednine
derived nine categories
ninecategories
categories (with
(withprimary
(withprimaryand
and
primary secondary
andsecondary
secondarytype).
type).(a)
(a)
type).
Fire
Fire
(a) and
and
Fire wood;
wood;
and (b)
wood; Earth
(b)(b)
Earth
Earthand
and Metal;
Metal;
and (c)
Metal; Earth
(c)(c)
Earth and
and
Earth Water;
Water;
and (d) Earth
(d)(d)
Water; Earth and
Earthand Wood;
Wood;
and (e)
Wood; Fire
(e)(e)
Fire and
and
Fire Metal;
Metal;
and (f)
Metal;(f)
Fire
Fire
(f) and
and
Fire Water;
Water;
and (g)
Water; Metal
(g)(g)
Metal
Metaland
and Water;
Water;
and Water;(h)
(h) Water
Water
(h) Waterand
and
andWood;
Wood;(i)
Wood; (i)Wood
(i) Woodand
Wood andMetal.
and Metal.
Metal.

3.2.
3.2.Overview
3.2. Overviewof
Overview ofofthe
theProposed
the ProposedSystem
Proposed System
System
The
Theproposed
The proposedreal-time
proposed real-timeiris-based
real-time iris-basedhealth
iris-based healthassessment
health assessmentsystem
assessment systemcontains
system containsvarious
contains variousstages
various stagesto
stages toperform
to perform
perform
human
human body
humanbody type
bodytype classification
typeclassification based
classificationbased on
basedonthe
onthe nine
thenine classes
nineclassesof
classesofbody
ofbodyconstitution.
bodyconstitution.The overall
constitution.The flowchart
Theoverall
overallflowchartof
flowchart
the
of proposed
ofthe
theproposedsystem
proposedsystem is shown
systemisisshown in
shownin Figure 6.
inFigure
Figure6. 6.
Symmetry 2020, 12,
Symmetry 2020, 12, 803
x FOR PEER REVIEW 99 of
of 17
17

Taking Iris
Image Pre-
Iris Segmentation
Processing
Image

Apply Training
Database CNN CNN Models

Training Process

Real time Classification of

Image
Shooting body constitution
Processing
Iris

Testing Process Determines potential organ

disorder with suggestions &
advice

Figure 6. Flowchart of the proposed iris-based health assessment system.

Figure 6. Flowchart of the proposed iris-based health assessment system.
3.2.1. Iris Image Acquisition
3.2.1.The
Iris proposed
Image Acquisition
system takes an iris image as input that is retrieved from an existing database or
captured by a handheld
The proposed high-resolution
system iris camera.
takes an iris image as input that is retrieved from an existing database or
captured by a handheld high-resolution iris camera.
3.2.2. Image Processing
3.2.2.In
Image Processing
the image processing stage, the color image is transformed into a grayscale image and
then Incontrast enhancement
the image processingisstage,
performed. The
the color color-to-grayscale
image is transformed transformation
into a grayscaleaims
imageto and
simplify
then
the
contrast enhancement is performed. The color-to-grayscale transformation aims to simplifythat,
information contained in the image and reduce the computational complexity. After the
contrast-limited adaptive
information contained in histogram
the image equalization (CLAHE)
and reduce the is applied
computational to further enhance
complexity. thecontrast-
After that, contrast
of the iris
limited image. histogram equalization (CLAHE) is applied to further enhance the contrast of the
adaptive
iris image.
3.2.3. Iris Segmentation
3.2.3.Iris
Irissegmentation
Segmentationis a difficult and challenging task because the anatomy of the iris is very complex.
Even iris and pupil boundaries are sometimes hard to distinguish. The purpose of iris segmentation
Iris segmentation is a difficult and challenging task because the anatomy of the iris is very
is to locate the boundary between the iris area and the sclera or the pupil. The detection of the iris
complex. Even iris and pupil boundaries are sometimes hard to distinguish. The purpose of iris
center is an essential section of the segmentation process because it contains useful information of
segmentation is to locate the boundary between the iris area and the sclera or the pupil. The detection
of the iris center is an essential section of the segmentation process because it contains useful
Symmetry 2020, 12, 803 10 of 17

each body organ and helps iridologists to examine the human health status. Different segmentation
methods and approaches have been proposed by researchers in the past. In this study, we applied
the iris segmentation method proposed in [33], which is a DL-based method. This iris segmentation
algorithm is a combination of learning-based and edge-based algorithms. The algorithm consists of
three phases that include eye detection, pupillary edge estimation, and limbus border assessment.
A well-trained faster R-CNN (Regions with CNN features) model is used to detect and classify the
position of the eye in a given image while a Gaussian mixture model (GMM) is used to find the
pupillary area. Then, the circular boundary of the pupil area is fitted based on the five key boundary
points. The limbus boundary points are located using a boundary point selection algorithm, and the
circular boundary of the limbus is created using these boundary points. In such a way, the iris area can
be accurately positioned. Interested readers can find more details of the iris segmentation algorithm
in [33].

3.2.4. Body Constitution Classification Using CNN Models

In this step, we performed experiments with several existing CNN models to classify the type of
body constitution using iris images. The models were trained based on the transfer learning technique.
We froze the parameters of most layers in the pre-trained CNN models, and only the parameters of the
last few layers were fine-tuned and retrained using annotated iris images. The classification accuracy
for the testing datasets was recorded to benchmark their performance. Figure 6 shows the flowchart of
the proposed real-time iris-based health assessment system.
For the practical system, it acquires the real-time iris image of the person with a handheld iris
camera, classifies the body constitution of the person with a trained CNN model, and displays the
classification results of the body constitution along with some useful advice for the subject to take care
of their health.

4. Experimental Results
This section presents the results of the experiment on the task of the iris-based health
assessment system.

4.1. Database
We used the CASIA-Iris-Thousand (a publicly accessible) database [34] to train the CNN models.
The database has been widely used in iris recognition research but has not yet been used in iridology.
It contains a total of 20,000 annotated iris images from 1000 subjects. All the iris images were annotated
(such as body constitution detail) and precisely labeled by experts from Taiwan International Institute
of Iridology (TIII).

4.2. Data Labeling and Partition

For the iris image annotation, we collaborated with experts from TIII. Iris segmentation and image
enhancement were performed on each image in the CASIA-Iris-Thousand database. Then, all the
segmented iris images were handed over to a few experts from TIII, who inspected each iris image and
gave a ground-truth label to each of them. The ground-truth label includes one of the nine categories
of body constitution as shown in Table 1. There were five experts who participated in performing this
annotation task. The labeling process consisted of two experts for an image. One expert gave the
category for one image and then the result of the label image was reviewed by the other expert, such as
a double-blind review.
The total number of iris images in the CASIA-Iris-Thousand database is 20,000. However, some of
them were removed because those images were not clear enough for the experts to read and give a
correct ground-truth label. In the end, the number of usable images was 5538. For data partition,
we partitioned all of the labeled data into 10 disjoint sets, which enabled us to perform 10-fold
cross-validation in the experiments. The number of images for each class is described in Table 5.
Symmetry 2020, 12, 803 11 of 17

Table 5. Number of images for each category.

Class Class Name Number of Images

0 Earth, Wood 616
1 Earth, Water 585
2 Earth, Metal 659
3 Wood, Metal 603
4 Water, Wood 598
5 Fire, Wood 611
6 Fire, Water 639
7 Fire, Metal 640
8 Metal, Water 587
Total 5538

4.3. Model Training

The pre-trained CNN models we tried included ResNet50 [21], InceptionV3 [22],
and DenseNet201 [23]. Each model was trained with a transfer learning technique. Our work mainly
used Keras and TensorFlow as deep learning frameworks to obtain pre-trained models, which were
originally trained on the ImageNet database [24].
To correctly assess the validity of the proposed method, the protocol of k-fold cross-validation
was used, which means that the experiments were repeated K times, and for each time, one partition of
data was held out as the test set and the other data partition was used as the training set to train the
model [35]. The model was tested using the test set and the accuracy was recorded. Such a procedure
was repeated k times until all data partitions were used as a test set. The final accuracy was reported
by calculating the mean of the k results from the folds. We set k=10 so that each partition contained an
approximately equal size of instances.

4.3.1. Inception V3
In 2014, Google scholar Szegdy et al. [22] proposed the inception network, which ranked first in
the classification and detection challenges in the 2014 ImageNet contest. Inception V3 is an inherited
and upgraded form of inception V1. The motivation of inception V3 is to bypass a demonstrative
obstruction, which means exceedingly decreasing the input size of the next layer. This is one of the
early adopters of the batch normalization features. It also uses factorization approaches for more
effective computations.

4.3.2. Residual Neural Network (ResNet)

The residual neural network (ResNet) was introduced by Kaiming He et al. [21]. ResNet is
the winner of the 2015-ILSVRC competition and became the first deeper network with 152 layers.
ResNet gained popularity through the use of skip connections and gives a technique for researchers to
design even deeper CNNs without affecting the generalization capabilities and accuracy. Furthermore,
it was among the foremost CNNs to use batch normalization characteristics.

4.3.3. Dense Convolutional Network (DenseNet)

Gao Huang et al. [23] proposed a novel convolutional network architecture, namely dense
convolutional network (DenseNet) to minimize the vanishing gradient problem. In DenseNet,
the feature map of every layer is directly connected to the next upcoming layer in a dense block.
DenseNet has various compelling benefits, such as it mitigates vanishing gradient problems, enhances
feature propagation, encourages feature reuse, and considerably reduces the number of parameters.
Compared with ResNet, DenseNet has more intermediate connections.
 Symmetry 2020, 12, x FOR PEER REVIEW 12 of 17
Symmetry 2020, 12, 803 12 of 17

4.4. Hyperparameters and Hardware

4.4. Hyperparameters and Hardware
In our experiments, a fixed learning rate of 0.01 was used with the Stochastics Gradient Descent
(SGD) optimizer
In our to train
experiments, all thelearning
a fixed networks.
rateAll networks
of 0.01 werewith
was used trained using 100 epochs
the Stochastics and
Gradient a batch
Descent
size of
(SGD) 64. The training
optimizer and
to train all thetesting stages
networks. Allutilized
networks eight Nvidia
were trainedTitan X 100
using (Pascal),
epochs two Intel
and Xeon
a batch E5-
size
2620 V4, and 128 GB Memory. We used Python as a programming language. Table 6 shows
of 64. The training and testing stages utilized eight Nvidia Titan X (Pascal), two Intel Xeon E5-2620 V4, the
hyperparameter
and 128 GB Memory. settings of the
We used experimented
Python CNN models.
as a programming language. Table 6 shows the hyperparameter
settings of the experimented CNN models.
Table 6. Hyperparameter settings of the experimented CNN models.
Table 6. Hyperparameter settings of the experimented CNN models.
Models Image Size Batch Size Epoch Learning Rate Optimizer Model Size
Inception
Models V3 Image
224 × 224
Size× 3 Batch Size
64 100 Learning
Epoch 0.01 SGD Model
Rate Optimizer 80.2MB
Size
ResNet50 229 × 229
Inception V3 224 × 224 × 3 × 3 6464 100
100 0.01
0.01 SGDSGD 87.6MB
80.2MB
DenseNet201
ResNet50 224× ×229
229 224× 3× 3 6464 100
100 0.01
0.01 SGDSGD 74.2MB
87.6MB
DenseNet201 224 × 224 × 3 64 100 0.01 SGD 74.2MB
4.5. Performance Analysis on Proposed Networks
4.5. Performance Analysis
This section on Proposed
analyzes Networks
the validity and performance of the proposed networks. Table 7 presents
the This
experimental results. the validity and performance of the proposed networks. Table 7 presents
section analyzes
the experimental results.
Table 7. Performance of the CNN models.
Table 7. Performance of the CNN models.
Models Mean Accuracy %
Models ResNet50 79.5
Mean Accuracy %
ResNet50 InceptionV3 77.5 79.5
InceptionV3 DenseNet201 82.9 77.5
DenseNet201 82.9
Table 7 shows the test accuracy of the three CNN models. The DenseNet201 network
outperforms the the
Table 7 shows others
test and achieves
accuracy of thethe highest
three accuracy,
CNN models. Thei.e., 82.9% compared
DenseNet201 networkwith the other
outperforms
models. The accuracy of the ResNet50 and InceptionV3 models is 79.5% and 77.5%, respectively.
the others and achieves the highest accuracy, i.e., 82.9% compared with the other models. The accuracy The
ofreceiver operation
the ResNet50 andcharacteristics (ROC) is
InceptionV3 models and areaand
79.5% under curve
77.5%, (AUC) are The
respectively. shown in Figure
receiver 7. The
operation
precision, recall, and F1 score bar graph is presented in Figure 8.
characteristics (ROC) and area under curve (AUC) are shown in Figure 7. The precision, recall, and F1
score bar graph is presented in Figure 8.

(a) (b) (c)

Figure
Figure7. The Receiver
7. The Operation
Receiver Characteristics
Operation (ROC) curve
Characteristics (ROC)of the experiment
curve of the results. (a) DenseNet201
experiment results. (a)
model, (b) InceptionV3 model, and (c) ResNet502 model.
DenseNet201 model, (b) InceptionV3 model, and (c) ResNet502 model.

Table 8 shows the performance of the ResNet50 model. In this table, we present the confusion
matrix, precision, recall, F1 scores, and average accuracy for each class. The ResNet model obtained a
79.5% total accuracy. The model performs well compared to inceptionV3.
Table 9 demonstrates the performance of the InceptionV3 model. In the table, we display the
confusion matrix, precision, recall, F1 scores, and average accuracy for each class. In the inceptionV3
model, we obtained a 77.5% accuracy. The performance of the model is not as good compared to
ReNet50 and Densenet201.
Symmetry
Symmetry2020, 12,12,
2020, 803x FOR PEER REVIEW 13ofof1717
13

(a) (b) (c)

Figure8. 8.
Figure Precision,
Precision, recall,
recall, andand F1 score
F1 score of allmodels
of all three three models in our experiments.
in our experiments. (a)(b)
(a) F1 score, F1Precision,
score, (b)
Precision,
(c) Recall. (c) Recall.

Table 8. Performance
Table 8 shows the performance analysis of model.
of the ResNet50 the ResNet 50 model.
In this table, we present the confusion
matrix, precision, recall, F1 scores, and average accuracy for
Predicted each class. The ResNet model obtained
Ground
a 79.5% total
Truth accuracy. The model performs well compared to inceptionV3.
0 1 2 3 4 5 6 7 8
0 571 Table38. Performance
10 2
analysis 4 ResNet
of the 14 50 model.
1 1 10
1 7 551 7 2 5 3 5 4 1
2 Ground Truth
26 10 514 8 Predicted
6 16 5 29 45
3 11 07 1 13 2 467 3 104 536 6 18 7 27 8 14
4 0 6 571
7 3 1 10 2 2 5394 1416 1 14 1 9 10 4
5 1 25 79 551 9 7 36 2 285 3368 5 50 4 74 1 12
6 2 2 263 10 2 514 6 8 96 1628 5 549 29 36 45 4
7 3 8 11
10 7 29 13 24 467 10
23 36 61 18 41 27 40814 36
8 4 13 6
13 7 26 1 15 2 539
9 16 17 14 22 9 33 4 439
Precision 50.853 25
0.898 90.841 9 0.83036 28
0.851 368
0.658 500.7787 74 0.65712 0.776
Recall 60.926 2
0.941 30.779 2 0.7746 9
0.901 28
0.602 549
0.859 36 0.6374 0.747
7 8 10 29 24 23 61 41
F1-Score 0.888 0.919 0.809 0.801 0.875 0.629 0.816 408 0.64736 0.762
8 13 13 26 15 9 17 22 33 439
Total Acc Precision 0.853 0.898 0.841 0.830 0.795
0.851 0.658 0.7787 0.657 0.776
Recall 0.926 0.941 0.779 0.774 0.901 0.602 0.859 0.637 0.747
F1-Score 0.888 0.919 0.809 0.801 0.875 0.629 0.816 0.647 0.762
Table 9. Performance analysis of the Inception V3 model.
Total Acc 0.795
Ground Predicted
Table 9 demonstrates
Truth 0 the1performance2 of 3the InceptionV3
4 5model. In
6 the table,
7 we display
8 the
confusion matrix, precision, recall, F1 scores, and average accuracy for each class. In the inceptionV3
0 558 5 11 7 5 19 3 3 5
model, we1obtained a977.5% 535accuracy.10The performance
6 3 of the9model is3 not as 4good compared
6 to
ReNet50 and
2 Densenet201.
13 13 527 22 8 11 1 28 36
3 5 4 13 464 11 40 9 31 26
4 4 Table 9.
3 Performance
5 12
analysis 17 V3 model.
528Inception
of the 9 15 5
5 15 15 10 40 27 361 48 76 19
6 Ground Truth
6 7 3 15 Predicted
11 32 523 27 15
7 5 012 1 42 2 34 3 164 5
75 6 40 7 375 8 41
8 0 17 558
12 5 33 11 22 7 125 19
19 3 18 3 29 5 425
Precision 1 0.882 9
0.882 5350.805 10 0.745 6 0.850 3 9
0.619 3
0.799 4 0.637 6 0.735
Recall 2 0.905 13
0.914 130.799 527 0.76922 0.882 8 11
0.5908 1
0.818 28 0.58536 0.724
3 5 4 13 464 11 40 9 31
F1-Score 0.894 0.898 0.8027 0.757 0.866 0.604 0.808 0.61026 0.729
4 4 3 5 12 528 17 9 15 5
Total Acc 5 15 15 10 40 0.775
27 361 48 76 19
6 6 7 3 15 11 32 523 27 15
7 5 12 42 34 16 75 40 375 41
Table 10 illustrate
8 the performance
17 12 of the33 DenseNet201
22 12 model. 19 The18
table contains
29 425the confusion
matrix, precision,Precision
recall, and F1 scores
0.882 and average
0.882 0.805 accuracy
0.745 0.850 for0.619
each class.
0.799 The overall
0.637 0.735performance
of the DenseNet201 Recall
model is 0.905 0.914 than
much better 0.799 ResNet50
0.769 0.882 0.5908 0.818 This
and InceptionV3. 0.585model
0.724 achieved an
F1-Score 0.894 0.898 0.8027 0.757 0.866 0.604 0.808 0.610 0.729
82.9% accuracy.
Total Acc 0.775
Symmetry 2020, 12, x FOR PEER REVIEW 14 of 17

Table 10 illustrate the performance of the DenseNet201 model. The table contains the confusion
matrix, precision,
Symmetry recall, and F1 scores and average accuracy for each class. The overall performance
2020, 12, 803 14 of 17
of the DenseNet201 model is much better than ResNet50 and InceptionV3. This model achieved an
82.9% accuracy.
Table 10. Performance analysis of the DenseNet201 model.

Ground Table 10. Performance analysis of the DenseNet201 model.

Predicted
Truth 0 1 2 3 Predicted
4 5 6 7 8
Ground Truth
0 1 2 3 4 5 6 7 8
0 587 1 8 2 2 6 4 2 4
0 587 1 8 2 2 6 4 2 4
1 2 562 7 1 3 6 2 1 1
1 2 562 7 1 3 6 2 1 1
2 13 6 564 15 4 9 3 19 26
2 13 6 564 15 4 9 3 19 26
3 10 6 17 485 9 24 15 22 15
3 10 6 17 485 9 24 15 22 15
4 6 3 2 2 556 8 10 6 5
4 6 3 2 2 556 8 10 6 5
5 5
15 15
6 6
1616 23
23 24
24 409 409 52 5256 10
56 10
6 6 2 2 0 0 11 66 6 6 30 30 562 56222 1022 10
7 7 8 8 13 13 4646 13
13 16 16 76 76 41 41398 398
29 29
8 8 9 9 9 9 3434 11
11 3 3 9 9 16 1625 25
471 471
Precision
Precision 0.900 0.9273
0.900 0.9273 0.811
0.811 0.869
0.869 0.892
0.892 0.7080.7080.797 0.797
0.722 0.722
0.824 0.824
RecallRecall 0.9520.9520.960 0.960 0.855
0.855 0.804 0.9297
0.804 0.9297 0.6690.6690.879 0.879
0.621 0.621
0.802 0.802
F1-Score 0.9250.9250.943
F1-Score 0.943 0.833
0.833 0.835 0.910
0.835 0.910 0.6880.6880.836 0.836
0.668 0.668
0.813 0.813
Total
Total Acc Acc 0.829
0.829

5. System Real-Time Demo

5. System Real-Time Demo
We developed a real-time body constitution classification system based on the proposed
We developed a real-time body constitution classification system based on the proposed method.
method. We used MATLAB R2019a to design a graphical user interface (GUI) of the system, which
We used MATLAB R2019a to design a graphical user interface (GUI) of the system, which is shown in
is shown in Figure 9.
Figure 9.

(a) (b)

(c) (d)
Figure
Figure9. 9.
The graphical
The user
graphical interface
user (GUI)
interface (GUI) demonstration
demonstration of of
thethe
health
health assessment
assessmentsystem. (a)(a)
system. TheThe
first window
first window of of
thethe
system. (b)(b)
system. Taking thethe
Taking irisiris
image. (c)(c)
image. Obtained
Obtained irisiris
image. (d)(d)
image. Health prediction.
Health prediction.

Because
Because this
this system
system waswas developed
developed in in
anan early
early stage
stage of of this
this study,
study, it it only
only classifies
classifies thethe body
body
constitution
constitution based
based ononthethe primary
primary attribute,
attribute, which
which means
means it performs
it performs five
five body-type
body-type classifications,
classifications,
instead
instead of of nine.
nine. Firstly,
Firstly, thethe system
system acquires
acquires anan input
input irisiris image
image captured
captured byby a handheld
a handheld irisiris camera.
camera.
Then,the
Then, thesystem
systemprocesses
processes ananiris
irisimage
imageusing
usinga deep learning
a deep algorithm
learning to find
algorithm tothe relationship
find between
the relationship
five elements
between and the and
five elements internal organs. At
the internal the final
organs. stage,
At the thestage,
final system performs
the system the body constitution
performs the body
classification, determines the possible most important body organ, and provides healthcare advice.
Symmetry 2020, 12, x FOR PEER REVIEW 15 of 17

constitution classification, determines the possible most important body organ, and provides
healthcare advice.
Symmetry 2020, 12, 803 15 of 17

Let us consider that the system identifies a person who belongs to the earth element. According
to the TCM principle,
Let us considerthe that
internal organs
the system related
identifies to earth
a person whoare the stomach
belongs to the earthand spleen.
element. The system
According to will
then perform a TCM diagnosis and display suggestions and advice to keep the body healthy.
the TCM principle, the internal organs related to earth are the stomach and spleen. The system will
then perform a TCM diagnosis and display suggestions and advice to keep the body healthy.
Health Assessment Results
Health Assessment Results
In this section, we present
In this section, thethe
we present health
healthassessment results
assessment results of of
thethe human
human bodybody constitution
constitution based based
on thefive-elements
on the TCM TCM five-elementstheorytheory
andand particularly important
particularly importantbody parts
body with with
parts characteristics and
characteristics and
suggestions. The results are shown in
suggestions. The results are shown in Figure 10. Figure 10.

(a) (b) (c)

(d) (e)
Figure 10. Health
Figure 10. assessment results
Health assessment of five
results of different types
five different ofofbody
types bodyconstitution according
constitution according to to
thethe TCM
TCM view. (a) Earth type body constitution. (b) Wood type body constitution.
view. (a) Earth type body constitution. (b) Wood type body constitution. (c) Water type body (c) Water type body
constitution. (d) Fire type body constitution. (e) Metal type body constitution.
constitution. (d) Fire type body constitution. (e) Metal type body constitution.
6. Conclusions and Future Work
6. Conclusions and
In this Future
paper, Work a non-invasive holistic iris-based human health assessment system
we proposed
based on the philosophy of TCM. This study combined the technique of iris image processing,
In this paper, we proposed a non-invasive holistic iris-based human health assessment system
TCM, and iridology to classify the human body constitution using a deep-learning-based algorithm.
based onFurthermore,
the philosophy of TCM.
a real-time operableThis study
system combined
based the technique
on the proposed of iris
method in this image
study processing, TCM,
was implemented
and iridology
to show to the classify theand
effectiveness human bodyofconstitution
practicability this study. The using a deep-learning-based
main advantage of the proposed system algorithm.
Furthermore,
is that itahelps
real-time operable
practitioners to make system
a quick based
analysis on
and the proposed
prompt method
classification in thisbody
of the subject’s study was
constitution and also gives instant feedback about how to maintain health. The
implemented to show the effectiveness and practicability of this study. The main advantage of the system is a user-friendly
and portable tool that can predict potential internal organ disorders and provide personal health care
proposed system is that it helps practitioners to make a quick analysis and prompt classification of
suggestions. The experimental results demonstrated that our iris-based method is effective and efficient
the subject’s
on thebody constitution and
CASIA-Iris-Thousand also The
database. gives instant
highest feedback
achieved about
accuracy howon
was 82.9% tothe
maintain health. The
DenseNet201
system isCNNa user-friendly
model. and portable tool that can predict potential internal organ disorders and
provide personal health care
In the future, we suggestions.
will The experimental
attempt to develop results demonstrated
new CNN architectures that will be able that ourthe
to learn iris-based
features of the nine classes more efficiently so that we can train our model
method is effective and efficient on the CASIA-Iris-Thousand database. The highest achieved with a lesser amount of
accuracy was 82.9% on the DenseNet201 CNN model.
In the future, we will attempt to develop new CNN architectures that will be able to learn the
features of the nine classes more efficiently so that we can train our model with a lesser amount of
data. In addition, we plan to publish our annotated database of the nine classes so that the research
Symmetry 2020, 12, 803 16 of 17

data. In addition, we plan to publish our annotated database of the nine classes so that the research
community of iridology can benefit from the labeled data.

Author Contributions: Conceptualization, Y.-H.L.; methodology, Y.-H.L.; software, C.-A.K. and S.-Y.T.; validation,
K.-L.Y.; formal analysis, K.-L.Y.; investigation, K.-L.Y.; resources, K.-L.Y.; data curation, K.-L.Y.; writing—original
draft preparation, M.S.A.; writing—review and editing, M.S.A.; supervision, Y.-H.L.; project administration,
Y.-H.L.; funding acquisition, Y.-H.L. All authors have read and agreed to the published version of the manuscript.
Funding: This research is funded by Ministry of Science and Technology in Taiwan under Contract no.
MOST 108-2221-E-008-071.
Acknowledgments: We give great appreciation to the Taiwan International Institute of Iridology (TIII) for helping
us labeling the Ground truth for iris images of CASIA-Iris-Thousand.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Ma, L.; Zhang, D.; Li, N.; Cai, Y.; Zuo, W.; Wang, K. Iris-based medical analysis by geometric deformation
features. IEEE J. Biomed. Health Inform. 2012, 17, 223–231. [PubMed]
2. Demea, A.L. Medical Diagnosis System based on iris analysis. Acta Tech. Napoc. Electron. Telecommun. 2009, 50.
3. Pushpalatha, M.; Muthurani, K. An Labeled Observations Iridology For Diagnosing Kidney Disease. Int. J.
Emerg. Technol. Comput. Sci. Electron. 2018, 25.
4. Othman, Z.; Prabuwono, A.S. Preliminary study on iris recognition system: Tissues of body organs in
iridology. In Proceedings of the 2010 IEEE EMBS Conference on Biomedical Engineering and Sciences
(IECBES), Kuala Lumpur, Malaysia, 30 November–2 December 2010; pp. 115–119.
5. Hussain, T.; Haider, A.; Muhammad, A.M.; Agha, A.; Khan, B.; Rashid, F.; Raza, M.S.; Din, M.; Khan, M.; Ullah, S.
An Iris based Lungs Pre-diagnostic System. In Proceedings of the 2019 2nd International Conference on Computing,
Mathematics and Engineering Technologies (iCoMET), Sukkur, Pakistan, 30–31 January 2019; pp. 1–5.
6. Hernandez, F.; Vega, R.; Tapia, F.; Morocho, D.; Fuertes, W. Early detection of Alzheimer’s using digital image
processing through iridology, an alternative method. In Proceedings of the 2018 13th Iberian Conference on
Information Systems and Technologies (CISTI), Cáceres, Spain, 13–16 June 2018; pp. 1–7.
7. Commons, W. File: Three Main Layers of the Eye.png—Wikimedia Commons, the Free Media Repository.
Available online: https://commons.wikimedia.org/w/index.php?title=File:Three_Main_Layers_of_the_Eye.
png&oldid=371960677 (accessed on 20 February 2020).
8. Lim, Y.W.; Park, Y.B.; Park, Y.J. A longitudinal study of iris parameters and their relationships with
temperament characteristics. Eur. J. Integr. Med. 2016, 8, 991–1000. [CrossRef]
9. Holl, R.M. Iridology: Another look. Altern. Health Pract. 1999, 5, 35–43. [CrossRef]
10. Um, J.Y.; An, N.H.; Yang, G.B.; Lee, G.M.; Cho, J.J.; Cho, J.W.; Hwang, W.J.; Chae, H.J.; Kim, H.R.; Hong, S.H.;
et al. Novel approach of molecular genetic understanding of iridology: Relationship between iris constitution and
angiotensin converting enzyme gene polymorphism. Am. J. Chin. Med. 2005, 33, 501–505. [CrossRef] [PubMed]
11. Lodin, A.; Demea, S. Design of an Iris-Based Medical Diagnosis System. In Proceedings of the International
Symposium on Signals, Circuits and Systems, Lasi, Romania, 9–10 July 2009; pp. 1–4.
12. Münstedt, K.; El-Safadi, S.; Brück, F.; Zygmunt, M.; Hackethal, A.; Tinneberg, H.-R.; Medicine, C. Can
iridology detect susceptibility to cancer? A prospective case-controlled study. J. Altern. Complement. Med.
2005, 11, 515–519. [CrossRef] [PubMed]
13. Hussein, S.E.; Hassan, O.A.; Granat, M.H. Control. Assessment of the potential iridology for diagnosing
kidney disease using wavelet analysis and neural networks. Biomed. Signal Process. Control 2013, 8, 534–541.
[CrossRef]
14. Ernst, E. Complementary/alternative medicine: Engulfed by postmodernism, anti-science and regressive
thinking. Br. J. Gen. Pract. 2009, 59, 298–301. [CrossRef]
15. Ernst, E. Iridology: Not useful and potentially harmful. Arch. Ophthalmol. 2000, 118, 120–121. [CrossRef]
[PubMed]
16. Zhao, C.; Li, G.-Z.; Wang, C.; Niu, J. Advances in patient classification for traditional Chinese medicine:
A machine learning perspective. Evid. Based Complement. Altern. Med. 2015, 2015, 376716. [CrossRef]
[PubMed]
Symmetry 2020, 12, 803 17 of 17

17. Chung, S.; Cha, S.; Lee, S.-Y.; Park, J.-H.; Lee, S. The five elements of the cell. Integr. Med. Res. 2017, 6, 452–456.
[CrossRef] [PubMed]
18. Worsley, J.R. Classical Five-Element Acupuncture: The Five Elements and the Officials; Redwing Book Company:
Damascus, OR, USA, 1998.
19. Irisology Talk. Available online: https://www.youtube.com/watch?v=oogtGRKxU2o&list=
PL8OTrtNdZxX31ArC5iMzSYaikP2POSqYF (accessed on 21 April 2020).
20. Krizhevsky, A.; Sutskever, I.; Hinton, G.E. Imagenet Classification with Deep Convolutional Neural Networks.
In Proceedings of the Advances in Neural Information Processing Systems, Las Vegas, NV, USA, 3–8 December
2012; Springer: Berlin/Heidelberg, Germany, 2012; pp. 1097–1105.
21. He, K.; Zhang, X.; Ren, S.; Sun, J. Deep Residual Learning for Image Recognition. In Proceedings of the IEEE
Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA, 27–30 June 2016; pp. 770–778.
22. Szegedy, C.; Vanhoucke, V.; Ioffe, S.; Shlens, J.; Wojna, Z. Rethinking the Inception Architecture for Computer
Vision. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV,
USA, 26 June–1 July 2016; pp. 2818–2826.
23. Huang, G.; Liu, Z.; Van Der Maaten, L.; Weinberger, K.Q. Densely Connected Convolutional Networks.
In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA,
21–26 July 2017; pp. 4700–4708.
24. Deng, J.; Dong, W.; Socher, R.; Li, L.-J.; Li, K.; Fei-Fei, L. Imagenet: A Large-Scale Hierarchical Image Database.
In Proceedings of the 2009 IEEE Conference on Computer Vision and Pattern Recognition, Miami, FL, USA,
20–25 June 2009; pp. 248–255.
25. Triwijayanti, A.; Suwastio, H.; Damayanti, R. Lung disorders detection based on irises image using
computational intelligent art. TEKTRIKA-Jurnal Penelitian dan Pengembangan Telekomunikasi, Kendali, Komputer,
Elektrik, dan Elektronika 2003, 8. [CrossRef]
26. Adelina, D.C.; Sigit, R.; Harsono, T.; Rochmad, M. Identification of diabetes in pancreatic organs using
iridology. In Proceedings of the 2017 International Electronics Symposium on Knowledge Creation and
Intelligent Computing (IES-KCIC), Surabaya, Indonesia, 26–27 September 2017; pp. 114–119.
27. Permatasari, L.I.; Novianty, A.; Purboyo, T.W. Heart disorder detection based on computerized iridology
using support vector machine. In Proceedings of the 2016 International Conference on Control,
Electronics, Renewable Energy and Communications (ICCEREC), Bandung, Indonesia, 13–15 September
2016; pp. 157–161.
28. Herlambang, R.A.N.P.; Isnanto, R.R.; Ajub, A.Z. Application of liver disease detection using iridology with
back-propagation neural network. In Proceedings of the 2015 2nd International Conference on Information
Technology, Computer, and Electrical Engineering (ICITACEE), Semarang, Indonesia, 16–18 October 2015;
pp. 123–127.
29. Miranda, J.D.; Salinas, S.A. Computational Measuring Approach for the Identification of Probable Intestinal
System Pathologies through the Human Iris Parameters. In Proceedings of the 2019 XXII Symposium on
Image, Signal Processing and Artificial Vision (STSIVA), Bucaramanga, Colombia, 24–26 April 2019; pp. 1–5.
30. Tang, H.; Huang, W.; Ma, J.; Liu, L. SWOT analysis and revelation in traditional Chinese medicine
internationalization. Chin. Med. 2018, 13, 5. [CrossRef] [PubMed]
31. Huan, E.-Y.; Wen, G.-H.; Zhang, S.-J.; Li, D.-Y.; Hu, Y.; Chang, T.-Y.; Wang, Q.; Huang, B.-L. Deep
convolutional neural networks for classifying body constitution based on face image. Comput. Math. Methods
Med. 2017, 2017. [CrossRef] [PubMed]
32. Wang, K.-C. The Five Elements Theory in Business Research. In Proceedings of the 46th Annual Meeting of
the Association of International Society for the Systems Sciences, Shanghai, China, 2–6 August 2002.
33. Li, Y.-H.; Huang, P.-J.; Juan, Y. An Efficient and Robust Iris Segmentation Algorithm Using Deep Learning.
Mob. Inf. Syst. 2019, 2019, 4568929. [CrossRef]
34. CASIA Iris Image Database. Available online: http://biometrics.idealtest.org/ (accessed on 17 January 2020).
35. Alpaydin, E. Introduction to Machine Learning; MIT Press: Cambridge, MA, USA, 2020.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).