Hindawi

Computational and Mathematical Methods in Medicine

Volume 2018, Article ID 8145713, 10 pages
https://doi.org/10.1155/2018/8145713

Research Article
Skin Disease Recognition Method Based on Image Color and
Texture Features

Li-sheng Wei , Quan Gan, and Tao Ji

School of Electrical and Engineer, Anhui Polytechnic University, Wuhu 241000, China

Correspondence should be addressed to Li-sheng Wei; [email protected]

Received 10 April 2018; Revised 24 July 2018; Accepted 31 July 2018; Published 26 August 2018

Academic Editor: John Mitchell

Copyright © 2018 Li-sheng Wei et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Skin diseases have a serious impact on people’s life and health. Current research proposes an efficient approach to identify singular
type of skin diseases. It is necessary to develop automatic methods in order to increase the accuracy of diagnosis for multitype skin
diseases. In this paper, three type skin diseases such as herpes, dermatitis, and psoriasis skin disease could be identified by a new
recognition method. Initially, skin images were preprocessed to remove noise and irrelevant background by filtering and
transformation. Then the method of grey-level co-occurrence matrix (GLCM) was introduced to segment images of skin disease.
The texture and color features of different skin disease images could be obtained accurately. Finally, by using the support vector
machine (SVM) classification method, three types of skin diseases were identified. The experimental results demonstrate the
effectiveness and feasibility of the proposed method.

1. Introduction subjective judgments, which may lead to misjudgments

and consequently delay the treatment of these. Therefore,
Composed of epidermis, dermis, and subcutaneous tissues, it is of great theoretical significance and practical value to
skin is the largest organ of human body, containing blood study how to extract symptoms of diverse skin diseases on
vessels, lymphatic vessels, nerves, and muscles, which can the basis of modern science and technology. Under this
perspire, perceive the external temperature, and protect circumstance, effective and accurate identification of the
the body. Covering the entire body, the skin can protect types of skin diseases can be achieved to prescribe treat-
multiple tissues and organs in the body from external ment according to patients’ symptoms [2].
invasions including artificial skin damage, chemical Over the past few years, the image processing technique
damage, adventitious viruses, and individuals’ immune has achieved rapid development in medicine. Some
system. Besides, skin can also avoid the loss of lipids to- equipment based on digital image technology has also been
gether with water within epidermis and dermis so that skin widely applied to people’s everyday life, for instance,
barrier function can be stabilized [1]. In spite of defense computed tomography (CT), digital subtraction angiog-
and barrier function, skin is not indestructible in that skin raphy (DSA), and magnetic resonance imaging (MRI).
tends to be constantly influenced by a variety of external Deeper research on this direction has been carried out by
and genetic factors. Currently, there are three main types scholars all over the world. For example, the skin disease
of skin diseases appearing in human body, including viral varicella was detected by Oyola and Arroyo [3] through
skin diseases, fungal skin diseases, and allergic skin dis- image processing technique’s color transformation,
ease. Despite the fact that these types of skin diseases can equalization as well as edge detection, and the image of
be cured at present, these diseases indeed have brought varicella was eventually collected and classified through
trouble to patients’ life. Nowadays, the majority of con- Hough transform. The final empirical results demonstrated
clusions on the patients’ existing symptoms are drawn that a better diagnosis was received in terms of detection
mainly based on doctors’ years of experience or their own on varicella, and preliminary test was also conducted on
2 Computational and Mathematical Methods in Medicine

varicella and herpes zoster on that basis. Chung and Sapiro cancer images across different types of neural network.
[4] put forward a method to detect the image of skin lesions Then, medical images were collected into this skin cancer
based on partial differential equation (PDE), with which classification system for training and testing based on the
a contour model of skin lesions was extracted on the basis matlab image processing toolbox.
of its morphological filtering through PDE. The final The above literature has made significant achievements
empirical results demonstrated that skin diseases could be on the identification of skin diseases. However, the proposed
accurately identified through this algorithm. Yu et al. [5] methods mainly aim at the identification of one type of skin
made a diagnosis on herpes simplex, varicella, and herpes disease, which makes them difficult to apply to the precise
zoster through reflectance confocal microscopy (RCM). identification of multitype skins. Therefore, in this paper,
The final empirical results demonstrated that specificity a method based on vertical image segmentation, GLCM, and
could be extracted from all the three different types of SVM is proposed to identify three various types of skin
herpes. Zhong et al. [6] diagnosed psoriasis vulgaris diseases, namely, herpes, paederus dermatitis, and psoriasis.
through three-dimensional computed tomography (CT) Firstly, the sample images of three skin diseases need to be
imageological technique of skin. The final empirical results preprocessed. Secondly, the vertical image is segmented and
demonstrated that diagnosis on psoriasis vulgaris of made corresponding geometric transformation. Based on
Munro’s microabscess was highly sensitive and specific. this, three types of skin diseases’ features are extracted, and
Sumithra et al. [7] proposed a novel approach for automatic their correlated parameters of feature texture and pixels of
segmentation and classification of skin lesions by using lesion areas are collected through image segmentation. Fi-
SVM and k-nearest neighbor (k-NN) classifier. Yasir et al. nally, the symptoms of herpes, paederus dermatitis, and
[8] proposed a detection system, which could be used in psoriasis are identified by utilizing the support vector ma-
computers or mobile phones, based on computer vision chine (SVM) method in order to improve identification
techniques. Arivazhagan et al. [9] proposed an automated accuracy.
system based on texture analysis for recognizing human The remaining structures of this paper are as follows.
skin diseases by independent component analysis of skin The Preliminaries section presents the identification
color image. The minimum distance classifier was used to method of skin diseases based on GLCM and SVM. The
classify the type of human skin diseases. Niu et al. [10] feature extraction and disease recognition are presented
carried out an experiment on the color image of skin er- in detail. Experimental results, comparisons, and dis-
ythema through HSV color space and fuzzy C-means cussions are provided in the Experiments and Analysis
clustering algorithm. The final empirical results demon- section, and finally conclusions are given in the Con-
strated that the time and accuracy of final segmentation clusions section.
could be improved through this method of dimensionality
reduction. Liu and Guo [11] diagnosed and identified skin 2. Preliminaries
tumors, vascular dermatosis, and psoriasis through com-
puted tomography (CT) based on the principle of optical The main steps of proposed methodology to skin disease
confocal, which has been widely applied. Luo et al. [12] recognition are shown in Figure 1. The flow chart contains
conducted diagnosis and identification on vitiligo through three phases: (1) processing of the original image; (2)
an independently developed analytic system of skin digital feature extraction; and (3) classification based on SVM. The
images. Both scientific and objective, the quantitative first stage is image processing. Because image may contain
evaluation method was worth studying. Lu et al. [13] some unwanted noise, it becomes necessary to filter the
classified smooth pixels by employing two-dimensional image to remove the noise. Then, by using image rotation
digital image segmentation and resizing combined with and segmentation, the representation of an image into
Markov random field (MRF), which demonstrated a reli- something that is more meaningful and easier to analyze.
able segmentation. The final results illustrated that the The specific areas of skin lesions are precisely divided, and
diagnosis and identification of psoriasis can be well solved. the identification accuracy is improved through vertical
Jaleel et al. [14] researched the skin cancer diagnosing image segmentation. The second stage is feature extraction,
methodology by using back propagation neural (BPN) in which the image texture features and color features of the
network classifier. Salimi et al. [15] presented the pattern skin lesions are further extracted. In texture feature ex-
recognition method to classify the skin disease. Ganesh- traction, GLCM is used to find the mathematical param-
kumar and Vasanthi [16] researched the melanoma disease eters of like contrast, correlation, entropy, uniformity, and
detection technique by using preprocessing and edge de- energy. The third stage is to identify the three various types
tection. Kolkur et al. [17] proposed a new human skin of skin diseases according to obtained features based on
detection algorithm to improve the recognition of skin SVM.
pixels, such as RGB (red, green, and blue), HSV (hue,
saturation, and value), and YCbCr (luminance and chro-
minance) color models. Gindhi et al. [18] and Kotian and 2.1. Image Prepocessing. Three common skin diseases are
Deepa [19] studied the problem of skin disease automated selected in this paper as the main research objects, which are
diagnosis system based on the techniques of image border herpes, paederus dermatitis, and psoriasis, respectively. Due
identification and feature data mining by matlab software. to differences in ways of acquiring the source images of skin
Kumar and Singh [20] established the relationship of skin epidermis and in resolution and size, image preprocessing
Computational and Mathematical Methods in Medicine 3

􏼌􏼌 􏼌 􏼌 􏼌
Image acquisition
􏼌􏼌 width 􏼌􏼌􏼌 􏼌􏼌􏼌 height 􏼌􏼌􏼌
widthnew � 2 × 􏼠􏼌􏼌􏼠􏼌 􏼌 􏼌
􏼡􏼌 + 􏼌􏼠 􏼡􏼌􏼌􏼡,
􏼌 2 × cos θ 􏼌􏼌 􏼌􏼌 2 × sin θ 􏼌􏼌
􏼌􏼌 􏼌 􏼌 􏼌 (2)
Image filtering 􏼌􏼌 height 􏼌􏼌􏼌 􏼌􏼌􏼌 width 􏼌􏼌􏼌
heightnew � 2 × 􏼠􏼌􏼌􏼠􏼌 􏼌 􏼌
􏼡􏼌 + 􏼌􏼠 􏼡􏼌􏼌􏼡,
􏼌 2 × cos θ 􏼌􏼌 􏼌􏼌 2 × sin θ 􏼌􏼌
Prepocessing Image rotation
where width and height each represent width and height of
the original image, while widthnew and heightnew refer to
Image segmentation width and height of the rotated image and θ is the angle
needing rotating.
The corresponding coordinates of image I and image Ic
Feature is as follows:
Texture feature Color feature
extration
x0 − xr1 􏼁 � x1 − xr2 􏼁 × cos θ − y1 − yr2 􏼁 × sin θ,
(3)
y0 − yr1 􏼁 � x1 − xr2 􏼁 × sin θ − y1 − yr2 􏼁 × cos θ,
Classification SVM classification
where x0 and y0 are the original coordinates, while x1 and y1
Figure 1: The identification process of skin diseases. are transformed ones; xr1 and yr1 are the central coordinates
of the original image, while xr2 and yr2 are the ones of
transformed image. Then, the image of neighborhood
on source images are needed for the sake of subsequent sampling transformation is processed via transformation of
vertical image segmentation. The specific processing is as Euclidean distance which is widely applied to the binary
follows. image and is particularly effective for the extraction of
Firstly, considering that noise constantly exerts skeleton. Accordingly, in this paper, the sample new
a negative impact on acquired samples of skin epidermis’s transformation image Ic is binarized first to obtain the
source images, it is necessary to denoise through median brightest pixel line, and the transformation of Euclidean
filtering for the reduction of the impact on skin seg- distance comes second. The specific process of the trans-
mentation and identification brought by irrelevant formation of Euclidean distance is as follows: through de-
background in the images. The median filter is a popular fining a binary image S under a two-dimensional (2D)
way to remove “salt-and-pepper” noise from an image and M × N, a background point set B(B ∈ S) is first assumed,
at the same time preserve edges and keep useful in- and then a foreground point set F(F ∈ S) meeting the
formation. In this paper, the median filter is adopted to premise B ∪ F � S is also assumed. Based on this, M is taken
preprocess and smoothen the source images. The used as a set in the medial axis transform meeting the premise
formula is as follows: M ∈ F. As a result, the shortest distance from a random
point px,y ∈ M to B can be calculated through Equation (4):
F′ x0 , y0 􏼁 � 􏼢 sort F(x, y)􏼣 , N ≥ 0, (1)
(x,y)∈S (N+1)/2
􏽱���������������
d􏼐px,y , bm,n 􏼑 � (x − m)2 +(y − n)2 , (4)
where F′ (x0 , y0 ) is the median of the gray value of the
image; S is the neighborhood collection of pixel; (x, y) is where d(•) represents the shortest Euclidean distance; px,y is
the element of S; F(x, y) is the gray value of (x, y); N is a random point in the medial axis transform set; x and y are
expressed as the number of the elements in the set of S; sort the coordinate values of the current point; bm,n is a point
is expressed as sequencing; and [g(•)](N+1)/2 is the median belonging to the background point set; and m and n are the
of the function g(•). The main idea of the median filter coordinate values of the current point. Binary mask image IBc
algorithm is to run through the signal gray value by gray is obtained from new image Ic of neighborhood sampling
value, replacing each gray value with the median of transformation, and IEc is eventually acquired via the
neighboring gray values. The pattern of neighbors is transformation of Euclidean distance. The brightest pixel
called the “window,” which slides, gray value by gray value, line can be obtained from the transformation of via the
over the entire signal. By using median filtering algorithm, application of IEc Euclidean distance, which is medial axis. In
the original images are denoised and their qualities are Figure 2(b), the image Ic is rotated to the horizontal co-
enhanced. ordinate. The binary mask image IBc is shown in Figure 2(c).
Secondly, denoise images are rotated in order to get the And, the Euclidean distance transform image IEc is presented
medial axis to segment the images. The denoised skin in Figure 2(d). In order to obtain the vertical segment image,
epidermis’s source images are processed via neighborhood the medial axis is marked out in Figure 2(d).
sampling with the intention of better obtaining the high- Finally, medial axis of the segmented region is located and
light line through the transformation of Euclidean distance. divided into ten vertical segments. It is necessary to create
The denoise image I is rotated to a specific angle θ, in which a bounding rectangle rbox on the image Ic in order to limit the
the width and height of new image Ic as shown in Equation image zone that is dotted with the white line. In Figure 3, it is
(2) via transformation of neighborhood sampling are shown that the medial axis is marked in light green after being
acquired: extracted from the transformation Euclidean distance.
4 Computational and Mathematical Methods in Medicine

Ic

(a) (b)

Medial axis

I Bc IEc

(c) (d)

Figure 2: Image of the transformation of Euclidean distance. (a) Filtering image. (b) Rotated image. (c) Binary mask image. (d) European
transformation image.

rbox rbox Lsum

Figure 3: The boundary rectangle of the image and medial axis. L1 L2 L3 L4 L5 L6 L7 L8 L9 L10

The perpendicular line for each point on the main axis can
be determined. So, the epithelium can be divided into ten Figure 4: Ten vertical image regions (L1 , L2 , . . . , L10 ).
vertical image regions, as done in [21]. The correspondingly
original medial axis is changed into ten short straight lines,
which is shown in Figure 4. After the ten vertical segment respectively. All the pictures as shown in Figure 5 are
images are determined, we will deal with these local images to extracted from photo gallery of sino-medicine website.
recognize the images L1 , L2 , . . . , L10 of skin disease.
2.2.1. Contrast
L−1 L−1
2.2. Texture Feature Extraction. Compared with the tradi- A1 � 􏽘 􏽘 (i − j)2 G(i, j), (5)
tional way, GLCM is an effective tool to analyze the features i j
of texture. The textures of different diseases in the skin
epithelial image can be obtained, such as contrast, corre- where |i − j| refers to the gray level difference between ad-
lation, entropy, uniformity, and energy. In this paper, three jacent pixels; G(i, j) is the distribution probability of gray-
common skin diseases are selected as the main research level difference between adjacent pixels; and A1 refers to the
objects, which are herpes, paederus dermatitis, and psoriasis, contrast, mainly used to describe the degree of depth of the
Computational and Mathematical Methods in Medicine 5

(a) (b)

(c) (d)

Figure 5: Types of skin diseases. (a) Normal skin. (b) Herpes. (c) Paederus dermatitis. (d) Psoriasis.

image textile grooves. The higher the contrast value goes, the vertical image segmentation. Then, we have the texture
deeper the grooves, and vice versa. Through experiment, we parameter of contrast in Figure 7.
can get the texture parameter of contrast of normal skin,
herpes, paederus, and psoriasis which is depicted in Figure 6.
2.2.3. Entropy
L−1 L−1

2.2.2. Correlation A3 � − 􏽘 􏽘 [G(i, j) · log G(i, j)], (7)

i j

L−1 L−1 where A3 refers to entropy, which measures the quantity of

􏽘i 􏽘j (i − x)(j − y)G(i, j)
A2 � , (6) information within the image and is changed with different
σxσy textures. As A3 increases, the textures of speck would be
arranged sparsely, and vice versa. When the entropy be-
where A2 refers to correlation, x is the mean value of the sum comes zero, there is no texture (Figure 8).
of the elements in each column in the square, y1 is the mean
value of the sum of the elements in each row, σ x the standard
deviation of the sum of the elements in each column, and σ y 2.2.4. Uniformity
the standard deviation of the sum of the elements in each L−1 L−1
1
row. Correlation is mainly used to describe the details of A4 � 􏽘 􏽘 G(i, j), (8)
relevant elements in each row and column in the process of i j (i − j)2 + 1
6 Computational and Mathematical Methods in Medicine

180 2

160 1.8

140 1.6
1.4
120
1.2
100
Contrast

Entropy
1
80
0.8
60
0.6
40 0.4
20 0.2
0 0
0 2 4 6 8 10 0 2 4 6 8 10
Number of vertical areas Number of vertical areas

Normal skin Paederus Normal skin Paederus

Herpes Psoriasis Herpes Psoriasis

Figure 6: Contrast. Figure 8: Entropy.

4 3.9
3.8
3.5 3.7
3.6
3 3.5
Correlation

Uniformity

3.4
2.5 3.3
3.2
2 3.1
3
1.5
0 2 4 6 8 10 0 2 4 6 8 10
Number of vertical areas Number of vertical areas

Normal skin Paederus Normal skin Paederus

Herpes Psoriasis Herpes Psoriasis
Figure 7: Correlation. Figure 9: Uniformity.

where A4 stands for the moment of inertia that is used for the From the above five figures, it can be seen that there are
description of the roughness of image texture. The larger the discrepancies in the discrete spots presented by different
value of A4 , the rougher the texture of endemic area will be, lesions. Table 1 shows the texture features of different skin
and vice versa (Figure 9). disease for the convenience of comparison.

2.2.5. Energy 2.3. Color Feature Extraction. In order to improve identi-

L−1 L−1
fication and treatment, the disease area of different skin
A5 � 􏽘 􏽘 G2 (i, j), diseases can be extracted on the basis of identifying the
(9)
i j texture features so that the area of herpes, paederus, and
psoriasis can be calculated, correspondingly. Given the fact
where A5 , the energy value mainly applied to describe the that most skin diseases appear in the shape of a circle, the
thickness of texture, is the quadratic sum of elements of watershed algorithm can be applied for image segmentation.
gray-level co-occurrence matrix in the horizontal and ver- In this paper, in order to avoid oversegmentation, marker-
tical directions. We can get the texture parameter of energy controlled watershed, combined with clustering, was per-
as it is depicted, respectively, in Figure 10. formed to segment the image. That is to say, to segment the
Computational and Mathematical Methods in Medicine 7

1500

1000

Energy
500

0
0 2 4 6 8 10
Number of vertical areas

Normal skin Paederus

Herpes Psoriasis
Figure 10: Energy.

Table 1: Texture features of different skin disease.

Contrast Correlation Entropy Uniformity Energy
Minimum Maximum Minimum Maximum Minimum Maximum Minimum Maximum Minimum Maximum
Normal skin 0.2053 0.3805 3.589 3.8917 1.0696 1.8224 3.8105 3.8974 843.7668 1437.6
Herpes 41.7036 178.2016 3.5308 3.7852 0.1355 0.5163 2.9878 3.6857 93.0058 345.2991
Paederus 51.676 87.3321 2.8167 3.2025 0.0741 0.1691 2.928 3.125 53.2181 121.5074
Psoriasis 8.713 40.3506 1.8612 3.6789 0.1849 0.5666 3.0617 3.4794 36.9206 205.6441

image by means of watershed algorithm that goes well with normal skin is nearly a straight line. Besides the pixel area
a proper threshold, cut each image into relatively small of normal skin, the pixel area of psoriasis is of the greatest,
zones, as Si , i � 1, 2, . . . , W. In the meantime, areas that herpes the second, and with that of paederus dermatitis
are mistakenly divided need to be highlighted. Then, it is being the lowest. These three diseases can be accurately
required to initialize the feature data of these areas such as identified through clear and distinct variation. However, if
Si , i � 1, 2, . . . , W into four clusters, namely, the normal some herpes once gets severe, the area of injured skin
skin zone, the skin disease zone, and the zone that is mis- could become larger. In this case, some certain errors may
takenly cut. Lastly, take each one in the four clustering forms emerge if we solely adhere to the statistics of pixel area.
as a large unit, among which each contains a smaller number Therefore, only by combining the features of texture with
of units. Then, look for one of the best smaller unit in the pixel area can the recognition rate of skin disease be
four larger units and take it as cluster center; continue to improved.
update the cluster center until the last center is no longer
changed. The results are shown in Figure 11.
Figure 11 shows the result of segmentation. Marker- 2.4. Classification Based on SVM. SVM is the statistical
controlled watershed segmentation that is fusion-based can method based on the statistical learning theory, which is
divide the lesion of skin disease in an effective way. Con- suitable for the classification of small sample numbers. It can
sidering there are still some black flecks which will have obtain the minimized training error and a confidence in-
a certain impact on the accuracy of experiments, in this terval term by analyzing the given training set to predict the
paper, morphological approach is utilized to process im- test set. In this paper, we use SVM to identify three different
ages, thus making the segmented image smoother so that it skin diseases. Firstly, the sample number and training
will be easier to count the black and white pixel in the number are selected from the extracted features (such as
certain area. In Figure 12, IO and IM of segmented area are color feature, and texture feature), and then by using the
picked up for inspecting after morphology processing. The rational kernel function of support vector machine, the
variation of different skin disease can be recognized by classification model can be established.
means of calculating the pixel area, as it is shown in In this paper, three common skin diseases herpes,
Figure 13. dermatitis, and psoriasis are represented as Class I, Class II,
In Figure 13, according to the statistics of these three and Class III, respectively. Texture feature classifier is the
kinds of skin diseases, it is shown that the pixel area of SVM1, lesion area feature classifier refers to SVM2, and
8 Computational and Mathematical Methods in Medicine

(a) (b) (c)

Figure 11: Image segmentation. (a) Original pictures. (b) Marker-controlled watershed segmentation. (c) Mark control + clustering
segmentation.

obtained from the characteristic model; yj refers to the re-

sults; and σ is a parameter value in the radial basis function.

3. Experiments and Analysis

To identify the classification, in this experiment, common
IO IM skin diseases (herpes, dermatitis, and psoriasis) are selected
as the research objects. Ninety images are also selected to be
identified accordingly, including herpes, dermatitis, and
psoriasis, thirty cases in each, together with twenty test
samples and ten standard samples. In this paper, the
combination of the color feature and the texture feature is
Figure 12: Morphological operation. employed to conduct the experiment, the results of which
are compared with those in [21, 22]. The comparison is
presented in Table 2.
integrated classifier is represented as SVM. On this Table 1 presents the results of different recognition al-
foundation, we change the different penalty factors to gorithms. It is used to identify images on twenty test samples
classify the maize diseases. As a result, the optimal rec- and ten standard samples of each skin disease. In the study
ognition effect can be realized when the penalty factor by Guo and Huo [22], the recognition accuracy is relatively
C � 50. The optimal classification function and the kernel low, with that of three diseases being at 75%, 80%, and 80%.
function are as follows: De et al. [21] mainly applies the method of unidirectional
⎪
⎧ ⎪
⎫ texture feature recognition, which can cause the relatively
⎨ n ⎬
⎠ + b∗ ,
⎝ 􏽘 a y k􏼐 x , x 􏼑 ⎞
f(x) � sgn⎪⎛ large errors. While in the study by De et al. [21], the method
⎩ j�1 j j i j ⎪
⎭ of support vector machine and genetic algorithm, with
(10) a rather ideal efficiency, are utilized for recognition, with the
�� �
��x − x ���2 recognition accuracy rate being 85%, 80%, and 85%. In this
⎜ � i j� ⎞
⎟ paper, the color feature and the texture feature can be used to
k􏼐xi , xj 􏼑 � exp⎛
⎝−
2
⎠,
2σ make up for the weakness of the unidirectional recognition
so that the recognition rate can reach to 90% and more,
where aj is a Lagrange multiplier, b∗ is the bias, and k(xi , xj ) which greatly improves the accuracy. Ten local vertical
is a kernel function; xi refers to the eigenvector that is images after segmentation can be obtained by the vertical
Computational and Mathematical Methods in Medicine 9

10000
9000
8000
7000

Skin area pixels

6000
5000
4000
3000
2000
1000
0
0 2 4 6 8 10
Number of vertical areas

Normal skin Paederus

Herpes Psoriasis
Figure 13: The calculation result of three skin diseases.

Table 2: The different methods to identify the results.

Herpes Dermatitis Psoriasis
Method Number Recognition Recognition Number Recognition Recognition Number Recognition Recognition
of test numbers rate (%) of test numbers rate (%) of test numbers rate (%)
Reference [22] 20 15 75 20 16 80 20 16 80
Reference [21] 20 17 85 20 16 80 20 17 85
The proposed
20 17 85 20 18 90 20 19 95
method

image segmentation, therefore improving the accuracy of Data Availability

skin diseases identification.
The data used to support the findings of this study are
available from the corresponding author upon request.
4. Conclusions
In this paper, the analysis method of vertical image seg- Conflicts of Interest
mentation is employed to identify three common skin
diseases. A number of irrelevant variables can be reduced The authors declare that they have no conflicts of interest.
through image filtering, image rotation, and Euclidean
distance transformation applied in image preprocessing. Acknowledgments
Then, the perpendicular line for each point on the main axis
can be determined. And, the epithelium can be divided into This study was funded by the National Natural Science
ten vertical image regions. Based on this, the grey-level co- Foundation of China (Grant 61203033), the Natural Science
occurrence matrix is adopted to extract the texture feature, Foundation of Anhui Province (Grant 1608085MF146), the
and the area pixel method is applied to extract the char- Natural Science Research Programme of Colleges and
acteristics of the lesion area. Finally, the support vector Universities of Anhui Province (Grant KJ2016A062), the
machine is utilized to classify the data of three different skin Visiting Study Foundation for Outstanding Young Talent of
diseases according to the features of the texture and the Anhui Educational Committee (Grant gxfxZD2016108), and
lesion area, achieving a more ideal accuracy of recognition. the Foundation for Talented Young People of Anhui
Nevertheless, the paper concentrating on herpes, dermatitis, Polytechnic University (Grant 2016BJRC008).
and psoriasis does not consider the different symptoms
caused by the same kind of skin disease. For instance, ec-
zema, herpes, and rubella all belong to the same series.
References
Therefore, it will be the focus of next step to recognize [1] Z. Hu and C. S. Yu, “Functional research and development of
different types of skin diseases of the same kind of series by skin barrier,” Chinese Journal of Clinicians, vol. 7, no. 7,
using image processing technique. pp. 3101–3103, 2013.
10 Computational and Mathematical Methods in Medicine

[2] V. Vezhnevets, V. Sazonov, and A. Andreeva, “A survey on processing and data mining,” International Journal of In-
pixel-based skin color detection techniques,” in Proceedings of novative Research in Computer and Communication Engi-
Graphicon 2003, pp. 85–92, Moscow, Russia, September 2003. neering, vol. 5, no. 4, pp. 8135–8141, 2017.
[3] J. Oyola and V. Arroyo, “Detection of chickenpox vesicles in [19] A. L. Kotian and K. Deepa, “Detection and classification of
digital images of skin lesions,” in Proceedings of 17th Iber- skin diseases by image analysis using MATLAB,” In-
oamerican Congress on Progress in Pattern Recognition, Image ternational Journal of Emerging Research in Management
Analysis, Computer Vision, and Applications, pp. 583–590, andTechnology, vol. 6, no. 5, pp. 779–784, 2017.
Buenos Aires, Argentina, September 2012. [20] S. Kumar and A. Singh, “Image processing for recognition of
[4] D. H. Chung and G. Sapiro, “Segmenting skin lesions with skin diseases,” International Journal of Computer Applica-
partial differential-equations-based image processing algo- tions, vol. 149, no. 3, pp. 37–40, 2016.
rithms,” IEEE Transactions on Medicine Image, vol. 19, no. 7, [21] S. De, R. Joe Stanley, C. Lu et al., “A fusion-based approach for
pp. 763–767, 2000. uterine cervical cancer histology image classification,” Com-
[5] S. R. Yu, X. H. Zhao, and X. M. Pu, “Image characteristics of puterized Medical Imaging and Graphics, vol. 37, no. 8,
dermopathic herpesvirus disease under reflectance confocal pp. 475–487, 2013.
microscope,” China Journal of Leprosy and Skin Diseases, [22] H. Guo and H. T. Huo, “Research on the application of gray
vol. 31, no. 2, pp. 85–88, 2015. level co-occurrence matrix for skin texture detection,” Journal
[6] L. S. Zhong, X. Jin, and C. Quan, “Diagnostic applicability of of Image and Graphic, vol. 15, no. 7, pp. 1074–1078, 2010.
confocal laser scanning microscopy in psoriasis vulgaris,”
Chinese Journal of Dermatovenereology, vol. 25, no. 8,
pp. 607-608, 2011.
[7] R. Sumithra, M. Suhilb, and D. S. Guruc, “Segmentation and
classification of skin lesions for disease diagnosis,” Procedia
Computer Science, vol. 45, pp. 76–85, 2015.
[8] R. Yasir, M. S. I. Nibir, and N. Ahmed, “A skin disease de-
tection system for financially unstable people in developing
countries,” Global Science and Technology Journal, vol. 3,
no. 1, pp. 77–93, 2015.
[9] S. Arivazhagan, R. N. Shebiah, K. Divya, and M. P. Subadevi,
“Skin disease classification by extracting independent com-
ponents,” Journal of Emerging Trends in Computing and In-
formation Sciences, vol. 3, no. 10, pp. 1379–1382, 2012.
[10] H. J. Niu, K. K. Shang, and Y. Liu, “Study of segmenting skin
erythema images by reducing dimensions of color space,”
Computer Engineering and Applications, vol. 13, no. 3,
pp. 219–221, 2006.
[11] F. Liu and H. X. Guo, “Research progress of CT in skin
diseases,” Chinese Journal of Dermatovenerol Integrated
Traditional Western Medicine, vol. 313, no. 3, pp. 189–191,
2014.
[12] W. Luo, R. S. Meng, and R. K. Cai, “Application of CDIA
system in measurement and research of vitiligo,” Chinese
Medical Journal, vol. 22, no. 12, pp. 1059-1060, 2011.
[13] J. Lu, E. Kazmierczak, and H. Jonathan, “Automatic seg-
mentation of scaling in 2-D psoriasis skin images,” IEEE
Transaction on Medical Imaging, vol. 32, no. 4, pp. 719–730,
2013.
[14] J. A. Jaleel, S. Salim, and R. B. Aswin, “Artificial neural
network based detection of skin cancer,” International Journal
of Advanced Research in Electrical, Electronics and In-
strumentation Engineering, vol. 1, no. 3, pp. 200–205, 2012.
[15] S. Salimi, M. S. Nobarian, and S. Rajebi, “Skin disease images
recognition based on classification methods,” International
Journal on Technical and Physical Problems of Engineering,
vol. 22, no. 7, pp. 78–85, 2015.
[16] M. Ganeshkumar and J. J. B. Vasanthi, “Skin disease iden-
tification using image segmentation,” International Journal of
Innovative Research in Computer and Communication Engi-
neering, vol. 5, no. 1, pp. 154–160, 2017.
[17] S. Kolkur, D. Kalbande, P. Shimpi, C. Bapat, and J. Jatakia,
“Human skin detection using RGB, HSV and YCbCr Color
models,” Advances in Intelligent Systems Research, vol. 137,
pp. 324–332, 2016.
[18] E. S. Gindhi, A. Nausheen, A. Zoya, and S. Ruhin, “An in-
novative approach for skin disease detection using image
 MEDIATORS of

INFLAMMATION

The Scientific Gastroenterology Journal of

World Journal
Hindawi Publishing Corporation
Research and Practice
Hindawi
Hindawi
Diabetes Research
Hindawi
Disease Markers
Hindawi
www.hindawi.com Volume 2018
http://www.hindawi.com
www.hindawi.com Volume 2018
2013 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Journal of International Journal of

Immunology Research
Hindawi
Endocrinology
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Submit your manuscripts at

www.hindawi.com

BioMed
PPAR Research
Hindawi
Research International
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Journal of
Obesity

Evidence-Based
Journal of Stem Cells Complementary and Journal of
Ophthalmology
Hindawi
International
Hindawi
Alternative Medicine
Hindawi Hindawi
Oncology
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2013

Parkinson’s
Disease

Computational and
Mathematical Methods
in Medicine
Behavioural
Neurology
AIDS
Research and Treatment
Oxidative Medicine and
Cellular Longevity
Hindawi Hindawi Hindawi Hindawi Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018