More Related Content
Similar to Detecting in situ melanoma using multi parameter extraction and neural classification mechanisms
Similar to Detecting in situ melanoma using multi parameter extraction and neural classification mechanisms (20)
More from IAEME Publication
More from IAEME Publication (20)
Detecting in situ melanoma using multi parameter extraction and neural classification mechanisms
- 1. INTERNATIONAL JOURNAL OF COMPUTER(IJCET), ISSN 0976 –
International Journal of Computer Engineering and Technology ENGINEERING
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1,(IJCET)
& TECHNOLOGY January- February (2013), © IAEME
ISSN 0976 – 6367(Print)
ISSN 0976 – 6375(Online)
Volume 4, Issue 1, January- February (2013), pp. 16-33
IJCET
© IAEME: www.iaeme.com/ijcet.asp
Journal Impact Factor (2012): 3.9580 (Calculated by GISI) ©IAEME
www.jifactor.com
DETECTING IN-SITU MELANOMA USING MULTI PARAMETER
EXTRACTION AND NEURAL CLASSIFICATION MECHANISMS
Ruksar Fatima1,Dr.Mohammed Zafar Ali Khan2,Dr. A. Govardhan3 and
Kashyap Dhruve4
1
(Head Dept. of Bio-Medical Engg, KBN College of Engg, Gulbarga, India,
Ruksar_blue@yahoo.co.in)
2
(Associate Professor,IIIT Hyderabad, Hyderabad, India,zafar @iith.ac.in)
3
(Director of Evaluation of JNTU, Hyderabad, India, dejntuh@jntuh.ac.in)
4
(Technical Director, Planet-i Technologies, Bangalore, India,kashyapdhruve@hotmail.com,)
ABSTRACT
Computer aided diagnostics systems are widely used for diagnosis of skin lesions.
aimed to enable early detection of skin cancer melanoma. The ܵܥܧܲܯadopts a supervised
This paper discusses a Multi Parameter Extraction and Classification System ( )ܵܥܧܲܯto
machine learning algorithm for classification. The dermoscopic images are represented by
operation of the ܵܥܧܲܯin the training and testing phase. The adoption of the Multilayer
extensive parameter sets extracted using a six phase approach. The paper discusses the
Feed Forward Neural Network ( )ܰܰܨܯclassifier is justified, its proficiency in accurately
diagnosing and classifying early signs of skin cancer melanoma in dermoscopic images of
skin lesions is proved through the experimental results discussed in the manuscript.
Keywords: Skin Cancer, Melanoma, Early Detection, In Situ Melanoma, MPECS , Multi
Parameter Extraction, Classification, Computer Aided Diagnostics, Image Processing, Skin
Lesions. Feature Extraction, Multilayer Feed Forward, Neural Network, Back Propagation,
SVM Classifier, ROC, Receiver Operating Characteristics
1. INTRODUCTION
Computer vision based diagnosis systems which are non-invasive have experienced a
wide acceptability ratio in the recent years. Deaths arising from skin cancers have increased
tremendously in the past decade [1] [2][3]. Skin Cancer Melanoma is dangerous and accounts
for maximum number of deaths arising due to skin cancers [3][4]. A recent survey [5]
conducted in 2012 for the United States of America alone, states that more than 75% of the
deaths arising from skin cancers are due to melanotic skin lesions. Skin cancer melanomas
16
- 2. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
are generally diagnosed by adopting biopsy procedures recommended by dermatologists.
Studies have proved that the mortality rate could be reduced if melanotic lesions are detected
The research work presented here puts forth the ܵܥܧܲܯto aid early detection of in
at a very early stage [6].
situ melanoma and classification of skin lesions. The dermoscopic images of skin lesions are
represented as a set of 21 features or parameter extracted phase wise. The skin lesions
represented as a set of features need to be classified. Statistical analysis carried out for on the
parameter set obtained post extraction were inconclusive in classifying the skin lesions [7]
into the .ܵܥܧܲܯMachine learning algorithms are adopted for accurate classification. A
hence there was an requirement for advanced classification mechanisms to be incorporated
ܰܰܨܯclassifier trained using the back propagation algorithm is adopted. The training phase
of the ܵܥܧܲܯenables the system to understand the features that exhibit properties of specific
classes defined during training. Subsequently the ܵܥܧܲܯcan be utilized for classification of
the testing data set. The proposed methodology is compared with the popular ܸܵ ܯclassifier.
This research manuscript is organized as follows. Section 2 discusses the
methodologies adopted by researchers. The next section discusses the proposed ܵܥܧܲܯat
literaturesurvey about skin cancer melanoma diagnosis systems and classification
length along with the ܰܰܨܯclassifier. The experimental evaluation and comparisons are
discussed in the penultimate section of this paper. The conclusion of the research work
presented in this paper is discussed in the fifth section.
2. LITERATURE SURVEY
The Asymmetry Border Color and Differential Structure (ABCD) methodology is the
most commonly used in the diagnosis of skin cancer melanoma [8]. To enhance the
efficiency additional optimizations such as fuzzy border extraction [9], “JSEG” based
optimization [10]; hybrid techniques [11] have been incorporated. A classification accuracy
of about 60% was achieved by incorporating wavelet decomposition techniques along with
the ABCD principle. Skin lesion classification can also be achieved using the Menzies
Method [12]. The ABCD based methods and the Menzies Method exhibit lesser accuracy in
detection of early stages of skin cancer melanoma. G. Betta at el [13] introduced a seven
point checklist method to diagnose skin lesions which proved to exhibit higher classification
accuracy than the other methodologies described. The computational complexity of the seven
point checklist method is considered as a major drawback. The, methods discussed above
exhibited less of low classification accuracy. The need of machine learning algorithms to
accurately classify and diagnose early symptoms of skin cancer melanoma arose. The use of
advance clustering techniques [14], k Nearest Neighborhood [6], classification/regression
trees [15] has been closely studied. The use of the support vector machines (ܸܵ )ܯfor
of skin cancer melanoma [18]. The ܵܥܧܲܯadopts the ܰܰܨܯbased classifier [23]. The
classification [16] [17] has been proposed by most of the researchers even for early detection
above mentioned methodologies are inaccurate to detect and classify early stages of skin
cancer melanoma. A few systems proposed to detect early symptoms exhibit low
classification accuracy as the early stages of skin cancer melanoma described through
features belong to the non-melanoma and melanoma classes cumulatively making
classification a very difficult or cumbersome task.
17
- 3. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
3. MULTI PARAMETER EXTRACTION AND CLASSIFICATION SYSTEM
(ࡹࡼࡱࡿ) MODELING
Appropriate skin lesions diagnosis at early stages enable possible cure to melanoma
skin cancers. Skin lesions of the melanotic type in the in situ stages are easily mistaken with
other skin related ailments by even experienced physicians. This paper introduces a computer
aided skin lesion diagnosis system named Multi Parameter Extraction and Classification
accounts for the highest number of deaths amongst skin cancers. The ܵܥܧܲܯrelies on the
System ( )ܵܥܧܲܯaiding early detection of skin cancer melanoma. Skin cancer melanoma
machine learning techniques incorporated to accurately classify the skin lesions defined by a
set of parameters extracted.
3.1. ࡹࡼࡱࡿ – PRELIMINARY NOTATION
Letܫௌ represent a set of dermoscopic skin lesion images for analysis defined as
ܫௌ ൌ ሼ݅ଵ , ݅ଶ , ݅ଷ , ݅ସ , … … ݅ ሽ
݅ represents an image
Where
݊represents the total number of images to be analyzed.
The MPECS system proposed in this paper is aids the early diagnosis of skin cancer
melanoma skin lesions. Let the classification set of the skin lesions be defined as
C ൌ ሼcଵ , cଶ , cଷ … cୡ୪ ሽ
Where c represents the class and cl represents the total number of classes considered.
Based on the above mentioned definitions it could be stated that i אI ! c אC | i אc.
The ܵܥܧܲܯaids the classification of skin lesions from dermoscopic images. The skin
lesions are defined as a set of parameter P that it exhibits. Let the Parameter set be defined by
P ൌ ሼpଵ , pଶ , pଷ , … pୟ ሽ
Where a represents the total number of p parameters
In the MPECS an image i אI is represented by a set of parameters P is used for
classification. The image I୬ could now defined as
I୬ ൌ ൛p୬ଵ , p୬ଶ , p୬ଷ , … p୬୮ ൟ
An image I୬ could be represented as a matrix and is defined as
18
- 4. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
x, ڮx,ୢ
I୬ ൌ ڭ ڰ ڭ൩
xୌ୲, ڮxୌ୲,ୢ
Where Ht represents the height and Wd represents the width of the image I୬ and x represents
the pixel value at the given location.
The MPECS adopts a 6 phase approach to extract the parameters set P i אIୗ . The
system considers 21 parameters for classification i.e. a ൌ 21. Each phase provides towards
the construction of the parameter set P.The ܵܥܧܲܯrelies on the supervised machine learning
with back propagation [24] training is adopted for classification. The trainings set ܶோ is
algorithm for classification. The use of multi-layer feed forward neural networks ()ܰܰܨܯ
defined as
ܶோ ൌ ሼ்ܴோଵ , ்ܴோଶ , ்ܴோଷ , … … . ்ܴோ ሽ
Where ்ܴோ represents the ݎ௧ training record.
c from the training image. I.e. ்ܴோ ൌ ሼI୰ c ሽwhere I୰ represents the parameter set that
The training record is constructed from the parameter set and the corresponding class
defines the ݎ௧ training image and c represents the corresponding classification class.
Similarly the test set ்ܶ is defined as
்ܶ ൌ ሼ்்ܴଵ , ்்ܴଶ , ்ܴோଷ , … … . ்்ܴ௧ ሽ
The objective of the ܵܥܧܲܯis to identify c௧ | ்்ܴ௧ אc௧ ܽ݊݀ c௧ ܥ א
3.2 OPERATION OF THE ࡹࡼࡱࡿ
The ܵܥܧܲܯrelies on supervised learning techniques for classification. To impart
intelligence the ܵܥܧܲܯis trained using the pre-classified dermoscopic images. The training
operation of the ܵܥܧܲܯis described by the algorithm given below where
்ܰܰܨܯோ ሺܺሻrepresents the back propagation training function based on the inputs set ܺ. The
parameters extracted from the phase is represented by ்ோ .
Algorithm Name:Training Phase of the ܵܥܧܲܯ
1. ݎtraining dermoscopic images
Input:
2. classification set C ൌ ሼcଵ , cଶ , cଷ … c୰ ሽ
1. Trained ܰܰܨܯclassifier.
Output:
19
- 5. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
Algorithm
Step 2: ܶோ ൌ
Step 1: Training Phase Begin
Step 3: For all training images ݅ ݎ ݐ 1
்ܴோ ൌ
Step 5: ࡹࡼࡱࡿ Parameter Extraction Begin
Step 4:
Parameter Set P்ோ ൌ
For phases :
Step 6:
P்ோ ൌ P்ோ ் ோ
Step 7:
Step 8:
Step 9: End phase
Step 11:்ܴோ ൌ P்ோ c
Step 10: Parameter Extraction End
Step 12: ܶோ ൌ ܶோ ்ܴ ோ
Step 14: Perform Training ்ܰܰܨܯோ ሺܶோ ሻ
Step 13: End For
Step 15: End Training Phase
The ܰܰܨܯpost training could be used for classification of the skin lesions represented in the
testing dermoscopic image dataset. The testing phase of the ܵܥܧܲܯis described in the algorithm
given below where்் represents the parameters extracted in the ௧ extraction phase
and ்்ܰܰܨܯሺܻሻis the classification function of ܻ defined as
்்ܰܰܨܯሺܻሻ ൌ ܿ௧
Where ܿ௧ ܥ א
Algorithm Name: Testing Phase of the ܵܥܧܲܯ
1. ݐtesting dermoscopic images
Input:
2. Trained ்ܰܰܨܯோ ሺܶோ ሻ
1. Classification Output ܿ௧ ܥ א
Output:
Algorithm
Step 2:்ܶ ൌ
Step 1: Testing Phase Begin
Step 3: For all testing images ݅ ݐ ݐ 1
்்ܴ ൌ
Step 5: ࡹࡼࡱࡿ Parameter Extraction Begin
Step 4:
Parameter Set P்் ൌ
For phases :
Step 6:
P்் ൌ P்் ்்
Step 7:
Step 8:
Step 9: End phase
Step 11:்்ܴ ൌ P்்
Step 10: Parameter Extraction End
Step 12: ்ܶ ൌ ்ܶ ்்ܴ
Step 14: Perform Classification Query ்்ܰܰܨܯሺ்ܶ ሻ ൌ ܿ௧
Step 13: End For
Step 15: End Testing Phase
20
- 6. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
The training and testing phases of the ܵܥܧܲܯhave been discussed above which
explains the operation. Parameter extraction is an integral part of the ܵܥܧܲܯwhich defines
ܵܥܧܲܯis acheied using a 6 phase approach described in the subsequent section of this paper.
the skin lesions and aids in training and classification. The 21 parameter extraction of the
3.3. PARAMETER EXTRACTION IN THE ࡹࡼࡱࡿ
The dermoscopic images of skin lesions is represented as a set of 21 parameters extracted
in 6 phases described below [7]
3.3.1. ܁۱۳۾ۻPARAMETER EXTRACTION – PHASE 1
present in an image ݅ ܫ אand extracting the Lesion Borders, the symmetry of the lesions and
This preliminary phase is primarily designed towards identifying the skin lesions
the color spreading factor of the skin lesions.
To extract the parameters discussed the images are resized using a bicubic
pixels in the nearest ሾ 4 ൈ 4 ሿproximity and is defined as
interpolation technique. The bicubic interpolation technique is an weighted average of the
ଷ ଷ
Pୖःऊ ሺx, yሻ ൌ ൫wt ୧୨ ൈ x୧ y ୨ ൯
୧
୧ୀଵ ୨ୀଵ
Where ݕ , ݔrepresent the pixel position and wt ୧୨is the weight at position ݅, ݆ of the image
i୬ אIୗ
Grey scaling is performed on the resized pixels, and the resultant grey scale pixel is
computed based on the following definition.
Pୋ୰ୣ୷ୗୡୟ୪ୣ ୧ ሺx, yሻ
ൌ ቂ0.2989 ൈ R ቀPୖःऊ ୧ ሺx, yሻቁቃ ቂ0.5870 ൈ G ቀPୖःऊ ୧ ሺx, yሻቁቃ
ቂ0.11400 ൈ B ቀPୖःऊ ୧ ሺx, yሻቁቃ
Where R ቀPୖःऊ ୧ ሺx, yሻቁrepresented the red channel value, G ቀPୖःऊ ୧ ሺx, yሻቁ represents the
green channel value and B ቀPୖःऊ ୧ ሺx, yሻቁrepresents the blue channel value of the resized
pixelPୖःऊ ୧ ሺx, yሻ.
Binirazation is performed on the grayscale image utilizing the adaptive thresholding
algorithm. The image noise elimination property of the adaptive thresholding based
binirazation algorithm is an additional parameter for adopting this algorithm. The adaptive
thresholding based binirazation is an iterative process wherein the thresholds are adapted
convergence is achieved. Let P୧୬ ୧ ሺx, yሻ represent the pixels of the binirized image i୬ . The
based on the regions it is being performed on. The iterative process is terminated when
regions of interest ሺ݅ݎሻ is extracted using the connected component labeling algorithm. The
ݏ’݅ݎobtained hold the information required and are identified by labels Lୖ୍ି୪ ୧ . The image
21
- 7. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
i୬ אIୗ based on the ܴܱ ݏ’ܫis defined as
i୬ ି୰୭୧ ൌ ሼL୰୭୧ିଵ ୧ L୰୭୧ିଶ ୧ L୰୭୧ିଷ ୧ ڮ L୰୭୧ି୪ ୧ ሽ
Where ݈represents the total number of ݏ’݅ݎextracted.
The centroids of the ݏ’݅ݎare computed based on the area enclosed by the .ݏ’݅ݎThe
centroids are computed to cluster the similar roi’s together in the same vicinity as it is
observed that the roi’s are similar in nature and can be clustered not effecting the
classification accuracy. This operation adopted enables roi’s exhibiting similar properties to
be clustered into a single cluster together represented asܴܱ .ܫThe labels L୰୭୧ି୪ ୧ used to
address the roi’s are clustered based on the energy computed and is defined as
ࣟgyୖ୍ି୪ ൫Lୖ୍ି୪ ୧ ൯ ൌ ∆൫L୰୭୧ିୟ ୧ , L୰୭୧ିୠ ୧ ൯
୮୭ୱ ୪ אୗ౨
Where ∆൫L୰୭୧ିୟ ୧ , L୰୭୧ିୠ ୧ ൯represents a function ∆൫L୰୭୧ିୟ ୧ , L୰୭୧ିୠ ୧ ൯ ൌ ሼ0,1ሽ i.e. if the
roi’s L୰୭୧ିୟ ୧ and L୰୭୧ିୠ ୧ are similar, 0 is returned and 1 in the case of dissimilarity. S୰୭୧ ൌ
ሼ1,2,3, … . lሽrepresents the set of roi’s obtained and a אS୰୭୧and b אS୰୭୧. pos represents the
position of the l୲୦ clustered ROI represented as Lୖ୍ି୪ ୧ .
The clustered ROI’s obtained define the image i୬ as
i୬ ିୖ୍ ൌ ሼLୖ୍ିଵ ୧ Lୖ୍ିଶ ୧ Lୖ୍ିଷ ୧ ڮ Lୖ୍ି୪ ୧ ሽ
Sobel edge detection is performed on the image i୬ ୖ୍. Post the edge detection the
distortion on the basis of the coordinate displacement and the frequency of distortion
observed is computed using the Fast Fourier Transforms. The axis based asymmetry
parameter of the lesion is obtained based on the principal component decomposition. The skin
lesion extracted is rotated such that the principal component coincides with the horizontal
p୶୧ୱ_ୗ୷୫୫ୣ୲୰୷ ൌ ቂ|i୬ ሺxሻ െ ıഥ ሺxሻ|ቃൗቂ i୬ ሺxሻቃ
axis. The skin lesion axis based asymmetry parameter is defined as
୬
Where i୬ ሺxሻthe original skin lesion and its reflected version is represented as ıഥ ሺxሻ.
୬
The symmetry parameter is computed for all the pixels within the lesion along the
based on the similarity of a pixel in a n ൈ n neighbourhood. The standard deviation of the
horizontal and vertical axis respectively. The color spreading factor of the lesions is computed
pixels in the neighborhood based on the Red channel, Green channel and Blue channel is
computed using
1
୬మ
തതതതതത
ൌ ඩ ଶ ሺChnl୧ െ Chnlሻଶ
n
σେ୦୬୪
୧ୀଵ
തതതതതത
Where Chnl ൌ ሼR େ୦୬୪ , Gେ୦୬୪ , Bେ୦୬୪ ሽ and Chnlrepresents the mean value defined as
22
- 8. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
1
୬మ
തതതതത
Chn ൌ ଶ Chn୧
n
୧ୀଵ
σେ୳୫୧୪୧୲୧୴ୣ ൌ σୖిౢ σୋిౢ σిౢ
The cumulative standard deviation for all the channels is defined as
ۍ ۍ ې ې
1 1
୬మ ୬మ
ൌ ێඩ ଶ ሺR େ୦୬୪ ୧ െ R େ୦୬୪ ሻଶ ۑ ێඩ ଶ ሺGେ୦୬୪ ୧ െ Gେ୦୬୪ ሻଶ ۑ
തതതതതതത തതതതതതത
ێn ୧ୀଵ ێ ۑn ୧ୀଵ ۑ
σେ୳୫୧୪୧୲୧୴ୣ
ۏ ۏ ے ے
ۍ ې
1
୬మ
ێඩ ሺBେ୦୬୪ െ Bେ୦୬୪ ሻଶ ۑ
തതതതതതത
ێnଶ ୧ୀଵ ୧
ۑ
ۏ ے
The color spreading factor of the lesion is computed using
pେ୪୰ିୗ୮୰ୣୟୢ ൌ 1 െ σ୭୰୫
Where σ୭୰୫is the normalized value of σେ୳୫୧୪୧୲୧୴ୣ between the range 0 and 1 and is obtained
σ୭୰୫ ൌ σେ୳୫୧୪୧୲୧୴ୣ ⁄σେ୳୫୧୪୧୲୧୴ୣିୟ୶
as
In order to extract the lesion border parameters the image is filtered using the
Gaussian and laplacian filter. The parameter describing the lesion borders is obtained by
computing the mean of the resultant pixels and is defined as
ୢ ୌ୲
pୣୱ୧୭୬ି୭୳୬ୢ୰୷ ൌ P୧ ሺx, yሻ൘ሾWd ൈ Htሿ
୶ୀଵ ୷ୀଵ
In this phase the parameters of the lesions such as the asymmetry along the horizontal
and vertical axis, the color spreading factor of the lesion and the boundary parameters is
discussed.
3.3.2. ܁۱۳۾ۻPARAMETER EXTRACTION – PHASE 2
The second phase of the ܵܥܧܲܯenables the extraction of the area, perimeter and the
eccentricity. This phase considers the binirized and clustered ܴܱ ܫimage as an input and
݅ ିோைூ ൌ ሼܮோைூିଵ ܮ ோைூିଶ ܮ ோைூିଷ ܮ ڮ ோைூି ሽ
computes the parameters based on the binirized image obtained from phase 1 defined as
Where ݈ represents the total number of ܴܱ ݏ’ܫidentified for the image ݅ .
The ܴܱ ݏ’ܫof image ݅ ିோைூ contain binirized pixel points i.e. black or white pixels.
The area parameter defines the white pixels that are encapsulated by the skin lesion and the
lesion area parameter is computed using
23
- 9. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
௧௦ିଵ
௦ି ൌ ሺ1⁄2ሻ | ൫ݔ ݕାଵ െ ݔାଵ ݕ ൯ |
,ୀ
Where ܲ݊ ݏݐis the number of points enclosed by the skin lesion defined by the ݈ ௧ ܴܱ . ܫAlso
݅, ݆ represent the ܲ݊ݏݐlocation and ൫ݔ , ݕ ൯ א൛൫ݔ , ݕ ൯ห ݂൫ݔ , ݕ ൯ ൌ 1 ሽ as ݔ , ݕ represents a
The perimeter defines the boundary ߂ ܤlength of the skin lesion. The lesion boundary
white pixel.
߂ܤሺ݈ሻ ൌ ሺܤሺ2: ݈, : ሻ െ ܤሺ1: ݈ െ 1, : ሻሻଶ
is computed using
Where ݈ represents the number of ܴܱ ݏ’ܫand its corresponding boundary is represented as .ܤ
The parameter defining boundary of the skin lesion is defined as
௦ି௨ௗ ൌ ඨ ߂ܤሺ݈ሻ
eccentricity is obtained by obtaining the moments represented as .ܯThe major axis is
The aspect ratio of the skin lesion is defined as the eccentricity of a skin lesion. The
ܣெ ൌ 2√2ሺඥܯ௫௫ ܯ௬௬ ܯ ሻ
defined as
ܣெ ൌ 2√2ሺඥܯ௫௫ ܯ௬௬ െ ܯ ሻ
The minor axis is defined as
Where the moments ܯ௫௫ , ܯ௬௬ , ܯ௫௬ andܯ are defined as
ܯ௫௫ ൌ ቂቀ ݔଶ ቁൗܰ 1⁄12ቃ
ܯ௬௬ ൌ ቂቀ ݕଶ ቁൗܰ 1⁄12ቃ
ܯ௫௬ ൌ ቂቀ ݕ כ ݔቁൗܰቃ
And
ܯ ൌ ටሺܯ௫௫ െ ܯ௬௬ ሻଶ 4ܯ௫௬ ଶ
The eccentricity parameter is computed using
௦ିா௧ ൌ ቈට൫ܣெ ଶ െ ܣெ ଶ ൯ൗܣெ
మ
3.3.3. ܁۱۳۾ۻPARAMETER EXTRACTION – PHASE 3
Researches have well understood the importance to 3 ܦdepth parameters to enhance the
classification accuracy [6] [19]. To obtain the 3 ܦdepth of the skin lesions the
ܵܥܧܲܯintroduced in this paper considers creating a 3 ൈ 3 masked windows represented as
ܹଵ ܹଶ ܹଷ
݅ ଷൈଷିெ௦ ൌ ܹସ ܹହ ܹ ൩
ܹ ଼ܹ ܹଽ
Where ܹ represent the weights of the pixel
The projection filter is defined as
24
- 10. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
ଽ
݅ ଷି௧ ൌ ܹ ܫ
ୀଵ
Where ܫ represents the intensity of the ݅ ௧ pixel.
The 3D depth projection parameter ௦ିଷ௧ is obtained by considering the geometric
௦ିଷ௧
mean defined as
ൌ ሾ1⁄9ሿ
ൈ ൣܫ௫ିଵ,௬ିଵ ܫ௫ିଵ,௬ଵ ܫ௫ିଵ,௬ାଵ ܫ௫,௬ିଵ ܫ௫,௬ ܫ௫,௬ାଵ ܫ௫ାଵ,௬ିଵ ܫ௫ାଵ,௬ଵ
ܫ௫ାଵ,௬ାଵ ሿ
Where ݕ ,ݔrepresent the horizontal and vertical pixel position.
3.3.4. ܁۱۳۾ۻPARAMETER EXTRACTION – PHASE 4
The fourth phase of the ܵܥܧܲܯis targeted towards obtaining the color components of the
skin lesions. The color components are extracted for the red channel blue channel and the
green channel. The mean and the variance parameters of each channel are considered as the
parameters to be utilized for classification. For the red channel the mean parameter is defined
∑ௐௗ ∑ு௧ ܴ ሺݕ ,ݔሻ
as follows
ோିெ ൌ
௫ୀଵ ௬ୀଵ
ܹ݀ ݐܪ כ
Where ݕ ,ݔrepresent the pixel positions.
The variance parameter is computed by obtaining the mean of all the ܤ columns , the
difference amongst them and then summed square difference divided by the height ݐܪ
∑ௐௗ ܴሺ݅ሻ
݊ܽ݁݉_݈݉ܥோ ሺ݅ሻ ൌ
ୀଵ
ܹ݀
݂݂݅ܦோ ሺ݅ ሻ ൌ ܴ݄݊ܥሺ: , ݅ ሻ െ ݊ܽ݁݉_݈݉ܥோ ሺ݅ሻ
∑ ݂݂݅ܦோ ሺ݅ሻ
ோି ൌ
ሺ ݐܪെ 1ሻ
∑ௐௗ ∑ு௧ ܩ ሺݕ ,ݔሻ
Similarly the green channel and blue channel parameters are defined as
ீିெ ൌ
௫ୀଵ ௬ୀଵ
ܹ݀ ݐܪ כ
∑ ீ݂݂݅ܦ ሺ݅ ሻ
ீି ൌ
ሺ ݐܪെ 1ሻ
∑ௐௗ ∑ு௧ ܤ ሺݕ ,ݔሻ
ିெ ൌ
௫ୀଵ ௬ୀଵ
ܹ݀ ݐܪ כ
∑ ݂݂݅ܦ ሺ݅ሻ
ି ൌ
ሺ ݐܪെ 1ሻ
25
- 11. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
3.3.5. ܁۱۳۾ۻPARAMETER EXTRACTION – PHASE 5
This phase of the ܵܥܧܲܯdiscusses the smoothening process of the ܴ ݈݄݊ܥand
ܩ using the 3 ൈ 3 masking procedure discussed in the third phase. The smoothing is not
adopted for the ܤ as the blue veils of the skin lesions is an important parameter for early
detection of skin cancer melanoma [20]. The smoothening filter for the ܴ ݈݄݊ܥand ܩ is
ଽ
defined as follows
݅ ோିௌ௧ ൌ ܹ ܴ݈݄݊ܥ
ୀଵ
ଽ
݅ ீ ൌ ܹ ܩ
ିௌ௧
ୀଵ
Where ܹ represent the weights of the pixel and ܴ݈݄݊ܥ , ܩ is red channel intensity and
green channel intensity of the ݅௧ pixel.
The red channel smoothened parameter ோିௌ௧ is defined as
ோିௌ௧
ൌ ሾ1⁄9ሿ
ൈ ൣܴ݈݄݊ܥ௫ିଵ,௬ିଵ ܴ݈݄݊ܥ௫ିଵ,௬ଵ ܴ݈݄݊ܥ௫ିଵ,௬ାଵ ܴ݈݄݊ܥ௫,௬ିଵ ܴ݈݄݊ܥ௫,௬
ܴ݈݄݊ܥ௫,௬ାଵ ܴ݈݄݊ܥ௫ାଵ,௬ିଵ ܴ݈݄݊ܥ௫ାଵ,௬ଵ ܴ݈݄݊ܥ௫ାଵ,௬ାଵ ሿ
The green channel smoothened parameter ீିௌ௧
ீିௌ௧ ൌ ሾ1⁄9ሿ
ൈ ቂܩ ௫ିଵ,௬ିଵ ܩ ௫ିଵ,௬ଵ ܩ ௫ିଵ,௬ାଵ ܩ ௫,௬ିଵ ܩ ௫,௬
ܩ ௫,௬ାଵ ܩ ௫ାଵ,௬ିଵ ܩ ௫ାଵ,௬ଵ ܩ ௫ାଵ,௬ାଵ ൧
3.3.6. ܁۱۳۾ۻPARAMETER EXTRACTION – PHASE 6
components of the skin lesion to obtain accurate classification results [21][22]. The ܵܥܧܲܯ
For early detection of skin cancer melanoma it is essential to extract all the color
discussed considers the cylindrical coordinate representation of pixels in the fourth and fifth
ܵܥܧܲܯconsiders the Cartesian representation of pixels. Hue, Saturation and Value
phase. In order to extract accurate and elaborate color parameters this phase of the
representation of the pixels are considered as the Cartesian representations. The mean and
from the ܴ , ܩ ܽ݊݀ܤ pixel values of the skin lesions. The hue value of a pixel is
variance parameters of the hue channel, variance channel and the value channel are extracted
43 ܩ| כ െ ܤ |
computed using the following definition
ۓቈ0 , ݈ܸܽݔܽܯൌ ܴ݈݄݊ܥ
ۖ ݈ܸܽݔܽܯെ ݈ܸܽ݊݅ܯ
ۖ
43 ܤ| כ െ ܴ|݈݄݊ܥ
݁ݑܪ, ൌ ቈ85 , ݈ܸܽݔܽܯൌ ܩ
۔ ݈ܸܽݔܽܯെ ݈ܸܽ݊݅ܯ
ۖቈ171 43 ݈݄݊ܥܴ| כെ ܩ |
ۖ
, ݈ܸܽݔܽܯൌ ܤ
ە ݈ܸܽݔܽܯെ ݈ܸܽ݊݅ܯ
Where ݈ܸܽݔܽܯൌ ݔܽܯሺܴܩ ,݈݄݊ܥ , ܤ ሻ is the maximum value of the red green and
blue channel of the pixel
26
- 12. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
݈ܸܽ݊݅ܯൌ ݊݅ܯሺܴܩ ,݈݄݊ܥ , ܤ ሻis the minimum value of the red green and blue channel
of the pixel
ሼ ݈ܸܽݔܽܯെ ݈ܸܽ݊݅ܯሽ
The Saturation and the value parameter of the pixel is computed using
ܵܽ݊݅ݐܽݎݑݐ, ൌ 255 כ
݈ܸܽݔܽܯ
ܸ݈ܽ݁ݑ, ൌ ݈ܸܽݔܽܯ
∑ௐௗ ∑ு௧ ݁ݑܪ ሺݕ ,ݔሻ
The mean of the hue channel and the is defined as
ு௨ ିெ ൌ
௫ୀଵ ௬ୀଵ
ܹ݀ ݐܪ כ
The variance is computed in a manner similar to the procedure described in phase 4 and is
∑ ݂݂݅ܦு௨ ሺ݅ ሻ
defined as
ு௨ ି ൌ
ሺ ݐܪെ 1ሻ
∑ௐௗ ∑ு௧ ܵܽ݊݅ݐܽݎݑݐ ሺݕ ,ݔሻ
The mean and the variance of the saturation channel is defined as
ௌ௧௨௧ ିெ ൌ
௫ୀଵ ௬ୀଵ
ܹ݀ ݐܪ כ
∑ ݂݂݅ܦௌ௧௨௧ ሺ݅ሻ
ௌ௧௨௧ ି ൌ
ሺ ݐܪെ 1ሻ
∑ௐௗ ∑ு௧ ܸ݈ܽ݁ݑ ሺݕ ,ݔሻ
Accordingly the Value channel parameters are obtained using the following equations
௨ ିெ ൌ
௫ୀଵ ௬ୀଵ
ܹ݀ ݐܪ כ
∑ ݂݂݅ܦ௨ ሺ݅ ሻ
௨ ି ൌ
ሺ ݐܪെ 1ሻ
The ܵܥܧܲܯdiscussed in this paper discusses the extraction 21 vital parameters which
would enable for classification of skin lesions especially targeted towards early detection of
skin cancer melanoma. The parameter extraction and the procedure involved in extraction are
achieved adopting a six phase approach discussed above.
3.4. ࡹࡲࡺࡺ CLASSIFIER
Figure 1 : A MFNN Classifier Structure
27
- 13. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
The classification of the testing data set of dermoscopic images requires a supervised
training procedure to be incorporated. A multilayer feed forward neural network ()ܰܰܨܯ
with the back propagation algorithm for training is adopted for the purpose of classification.
of the neural network. A sample ܰܰܨܯis shown in Fig 1. The ܰܰܨܯconsists of a set of
The parameters extracted from the testing image set are directly provided to the input neurons
input neurons , ܯnumber of hidden layers and an output layer.The ܰܰܨܯclassifier is a
machine learning algorithm. The learning and classification is achieved based on the
layer represented as ܽఊ where ߛ ൌ 1,2,3, … c୰ and c୰ is the trained ݎclasses.
weightsࣱ which are changed in the neighboring layers. The input to the neuron in every
ࣿࣾషభ
The output of the neuron is defined as
ܽࣼ ൌ ݂൫߮ࣼ ൯ ൌ ݂ ቌ ࣱࣼࣻ ܽࣻ
ሺࣾିଵሻఊ
ቍ
ࣾఊ ࣾఊ ࣾ
ࣻୀ
݉represents the index of the hidden layer i.e. 1 ݉ ܯ
Where
݊ represents the number of neurons in݉௧ layer
݆ represents the index ݆ | 1 ݆ ݊
߮ ࣼ represents the weighted sum of input values for ࣼ ௧ neuron in layer ܽ
ࣱࣼ represents the weight between ࣼ ௧ neuron in ݉௧ layer and ݅ ௧ neuron in ሺ݉ െ 1ሻ௧ layer
γ
The classification error ߜࣼ ം observed at the output layer is obtained from the following
ࣧ
definition
ܽࣼ ൌ ݂ ′ ൫߮ࣼ ൯ߜࣼ ൌ ݂ ′ ൫߮ࣼ ൯൫ࣼݕ െ ࣼݕ൯
ࣾఊ ெఊ ఊ ெఊ ࣴఊ ఊ
The sum of square errors observed through the ܰܰܨܯis defined as
1
ߦఊ ൌ ൫ߜ ൯
ఊ ଶ
2
ࣼୀଵ
ࣾశభ
The error existent at the output layers is propagated back and is defined as
ൌ݂ ൫߮ࣼ ൯ ߜℓ
ሺࣾାଵሻఊ ሺࣾାଵሻ
ߜࣼ ࣱℓࣼ
ࣾఊ ᇱ ࣾఊ
ℓୀଵ
The errors propagated is minimized by altering the weights of the neighboring neurons and
∆ஓ ࣱ ൌ ߟߜࣼ ࣻݑ
ࣾఊ ሺࣾିଵሻఊ
the weight update function is defined as
ୡ౨ ୡ౨
1
The classification error that exist through
ଶ
ߦ௦௦ ൌ ߦఓ ൌ ቀ ࣼݕെ ߮ ࣼ ቁ
ஓ ஓ
2
ఓୀଵ ఓୀଵ ࣼୀଵ
The ܰܰܨܯClassifier is designed to minimize the classification error by effective
weight adjustments achieved due to the back propagation algorithm.
28
- 14. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
The ܵܥܧܲܯadopts machine learning techniques for classification. The ܰܰܨܯbased
classifier is adopted for classification which classifies based on the supervised learning
imparted using the back propagation algorithm. Parameter extraction is achieved using a six
phase approach to define the skin lesion. The training dataset constitutes of a set of
dermoscopic images which are pre-classified with the aid of dermatologists. The images
set. The ܰܰܨܯclassifier is trained using the training data set. The testing set consists of
defined in terms of the parameters and the classified type constitute towards the training data
dermoscopic images that need to be diagnosed. The diagnostic results are represented as
using the ܵܥܧܲܯparameter extraction phases. The resulting dataset is termed as the test
classes. To enable diagnosis the test images are represented as a set of parameters extracted
dataset. The test dataset is provided to the trained ܰܰܨܯfor classification. The subsequent
ܸܵ ܯclassifiers.
section of this paper discusses the experimental verification and comparisons with the
4. EXPERIMENTAL STUDY AND COMPARISONS
The ܵܥܧܲܯdiscussed in this paper was realized on the MATLAB platform. In order to
evaluate the ܵܥܧܲܯtraining and testing dataset are to be created. To construct the training
and testing datasets dermoscopic images are obtained from the “The Atlas of Dermoscopy”
[25]. The atlas considered is a collection of over 2000 skin lesion images obtained from
a reference to evaluate the classification accuracy of the ܵܥܧܲܯsystem proposed. The use of
various patients. The atlas also provides the classification and diagnosis data which is used as
the atlas is charted as others researchers too have used the same atlas to evaluate their
ܵܥܧܲܯproposed is compared with the popular ܸܵ ܯclassifiers as proposed by Xiaojing
proposed diagnosis mechanisms especially skin cancer melanoma [20] [26]. The
Yuan et al.[18].
CLASSIFICATION ERROR
14 13.33 %
12
CLASSIFICATION ERROR IN %
10
8
6
4
2
< 1%
0
MEPCS - CLASSIFICATION ERROR SVM CLASSIFIER - CLASSIFICATION ERROR
Figure 2 : Classification Error Observed
29
- 15. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
The training dataset consists of pre-classified dermoscopic images. We have considered a
set of 45 dermoscopic images from the “Atlas of Dermoscopy” as the training data set. The
training dataset consist of skin lesion images of primarily 3 classes namely “Advanced
Melanoma”, “Early Melanoma” and “Non-Melanoma”. The 21 parameters extracted per
image using the 6 phase approach contribute to each record of the training data set along with
The ܵܥܧܲܯand the ܸܵ ܯClassifier was trained using similar training datasets. Similar test
the classified name. A similar approach is followed to construct the testing dataset.
data sets were evaluated used the ܵܥܧܲܯand ܸܵ ܯclassifier. The classification results
obtained are provided to the Receiver Operating Characteristic (ܴܱܥሻ analysis tool. The error
classification error of the ܸܵ ܯclassifier is greater than the error exhibited by the proposed
in classification obtained is graphically shown in Fig 2. From the figure it is evidentthat the
ܵܥܧܲܯsystem.The ܵܥܧܲܯand the ܸܵ ܯclassifier classifiers are compared on the basis of
the receiver operating characteristic (ܴܱ )ܥanalysis. The ܴܱ ܥanalysis is carried out using a
The ܴܱ ܥanalysis plot obtained is shown in Fig 3.
tool developed using C#.Net programming language on the Visual Studio 2010 platform.
Receiver Operating Characteristic (ROC) - CURVE
1
0.9
0.8
0.7
SENSITIVITY
0.6
0.5
0.4
0.3
0.2 MPECS SVM CLASSIFIER
0.1
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Figure 3 : ۱۽܀Curve Analysis
1- SPECIFICITY
The classification accuracy and the efficiency of the ܵܥܧܲܯClassifier along with the
ܸܵ ܯClassifier are diagrammatically shown in Fig 4 and Fig 5. The ܵܥܧܲܯsystemexhibits
an 81% accuracy in classification as compared to the 75% accuracy exhibited by the ܸܵܯ
Classifier. The ܵܥܧܲܯsystem proposed in this paper improves the efficiency in
classification by about 7.4% when compared to the ܸܵ ܯclassifier.The experimental study
the ܵܥܧܲܯsystem proposed in this paper is capable of accurately classifying skin lesions
conducted and from the results discussed in this section of the paper, it can be concluded that
exhibiting early melanotic symptoms when compared to the popularly used ܸܵ ܯclassifier
systems.
30
- 16. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
CLASSIFIER ACCURACY
0.81
0.8
0.79
0.78
ACCURACY
0.77
0.76
0.75
0.74
0.73
0.72
MEPCS ACCURACY SVM CLASSIFIER ACCURACY
Figure 4 : Classification Accuracy Comparisons
0.83 CLASSIFIER EFFICIENCY
0.82
0.81
0.8
EFFECIENCY
0.79
0.78
0.77
0.76
0.75
0.74
0.73
MPECS EFFICIENCY SVM CLASSIFIER EFFICIENCY
Figure 5: Classification Efficiency Comparisons
5. CONCLUSION
This paper highlights the benefits and the requirement for early detection of skin cancer
paper relies on a supervised ܰܰܨܯbased classifier to accurately classify skin lesions and enable
melanoma. The Multi Parameter Extraction and Classification System ( )ܵܥܧܲܯdiscussed in this
early detection of skin cancer melanoma. The ܵܥܧܲܯcomprehensively defines skin lesions in terms
of the ܵܥܧܲܯis discussed in this paper. The ܵܥܧܲܯsystem is trained using the Back Propagation
of a parameter sets extracted using the six phase approach. The training and testing phase algorithms
Algorithm. The trained classifier of the ܵܥܧܲܯis used for analysis and diagnosis of skin lesions on
unknown types. The classification preeminence of the ܵܥܧܲܯis proved over the popular SVM
classifier systems.The results obtained and discussed prove that the ܵܥܧܲܯsystem proposed in this
paper aids the detection of skin cancer melanoma in the naïve stages thereby improving computer
aided diagnosis mechanisms and early treatment initiatives to be adopted by dermatologists for the
benefit of patients suffering from the life threatening melanotic skin cancer diseases.
31
- 17. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
REFERENCES
[1] R. Marks, “Epidemiology of melanoma,” Clin. Exp. Dermatol., vol. 25, pp. 459–463,
2000
[2] World Health Organization, Ultraviolet Radiation and the INTERSUN Programme.
(2007. Dec.). [Online]. Available: http://www.who.int/uv/faq/skincancer/en/
[3] Imran Ali, Waseem A. Wani and KishwarSaleem,"Cancer Scenario in India with Future
Perspectives",Cancer Therapy Vol 8, 56-70, 2011
[4] SinhaR,Anderson DE, Mc Donald SS and Greenwald P, "Cancer Risk and Diet in India",J
Postgrad Med 2003,49: 222-228
[5] American Cancer Society,"Cancer Facts and Figures 2012". 2012, American Cancer
Society, Inc. pp 1-68
[6] Brian D’Alessandro, Atam P. Dhawan,"3D Volume Reconstruction of Skin Lesions for
Melanin and Blood Volume Estimation and Lesion Severity Analysis",IEEE Transactions on
Medical Imaging, July 2012 DOI : 10.1109/TMI.2012.2209434
[7] Ruksar Fatima, Mohammed Zafar Ali Khan, A. Govardhan and Kashyap D
Dhruve."Computer Aided Multi-Parameter Extraction System to Aid Early Detection of Skin
Cancer Melanoma ",IJCSNS International Journal of Computer Science and Network
Security,Vol. 12 No. 10 pp. 74-86.2012
[8] G. Argenziano, H. P. Soyer, S. Chimenti, R. Talamini, R. Corona, F. Sera, M. Binder, L.
Cerroni, G. De Rosa, G. Ferrara, R. Hofmann-Wellenhof,M. Landthaler, S.W.Menzies, H.
Pehamberger, D. Piccolo, H. S. Rabinovitz, R. Schiffner, S. Staibano, W. Stolz, I. Bartenjev,
A. lum, R. Braun, H. Cabo, P. Carli, V. De Giorgi, M. G. Fleming, J. M. Grichnik, C. M.
Grin, A. C. Halpern, R. Johr, B. Katz, R. O. Kenet, H. Kittler, J. Kreusch, J. Malvehy, G.
Mazzocchetti, M. Oliviero, F. Ozdemir, K. Peris, R. Perotti, A. Perusquia, M. A. Pizzichetta,
S. Puig, B. Rao, P. Rubegni, T. Saida, M. Scalvenzi, S. Seidenari, I. Stanganelli, M. Tanaka,
K. Westerhoff, I. H. Wolf, O. Braun-Falco, H. Kerl, T. Nishikawa, K. Wolff, and A. W. Kopf,
“Dermoscopy of pigmented skin lesions: Results of a consensus meeting via the Internet,” J.
Amer. Acad. Dermatol., vol. 48, no. 5, pp. 680–693, 2003.
[9] V. T. Y. Ng, B. Y. M. Fung, and T. K. Lee, “Determining the asymmetry of skin lesion
with fuzzy borders,” Comput. Biol. Med., vol. 35, pp. 103–120, 2005
[10] M. E. Celebi, Y. A. Aslandogan,W. V. Stoecker, H. Iyatomi, H. Oka, and X. Chen,
“Unsupervised border detection in dermoscopy images,” Skin Res. Technol., vol. 13, no. 4,
pp. 454–462, 2007
[11] S. E. Umbaugh, R. H. Moss, and W. V. Stoecker, “An automatic color segmentation
algorithm with application to identification of skin tumor borders,” Comput. Med. Imag.
Graph., vol. 16, pp. 227–235, 1992
[12] V. T. Y. Ng, B. Y. M. Fung, and T. K. Lee, “Determining the asymmetry of skin lesion
with fuzzy borders,” Comput. Biol. Med., vol. 35, pp. 103–120, 2005
[13] G. Betta, G. Di Leo, G. Fabbrocini, A. Paolillo, and M. Scalvenzi, “Automated
application of the “7-point checklist” diagnosis method for skin lesions: Estimation of
chromatic and shape parameters,” in Proc. Instrum. Meas. Technol. Conf. (IMTC 2005), May,
vol. 3, pp. 1818– 1822.
[14] Tasoulis, S.K.; Doukas, C.N.; Maglogiannis, I.; Plagianakos, V.P.; , "Classification of
dermatological images using advanced clustering techniques," Engineering in Medicine and
Biology Society (EMBC), 2010 Annual International Conference of the IEEE , vol., no.,
pp.6721-6724, Aug. 31 2010-Sept. 4 2010 doi: 10.1109/IEMBS.2010.5626242
32
- 18. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 1, January- February (2013), © IAEME
[15] Tim K. Lee*, M. Stella Atkins, Michael A. King, Savio Lau, and David I.
McLean´Counting Moles Automatically From Back Images‖ IEEE TRANSACTIONS ON
BIOMEDICAL ENGINEERING, VOL. 52, NO. 11, NOVEMBER 2005
[16] Faal, M., MiranBaygi, M. H. and Kabir, E. (2013), Improving the diagnostic accuracy
of dysplastic and melanoma lesions using the decision template combination method. Skin
Research and Technology, 19: e113–e122. doi: 10.1111/j.1600-0846.2012.00617.x
[17] M. E. Celebi, H. A. Kingravi, B. Uddin, H. Iyatomi, Y. A. Aslandogan, W. V.
Stoecker, and R. H. Moss, ―A methodological approach to the classification of dermoscopy
images,‖Comput. Med. Imag. Graph., vol. 31, pp. 362–373, Mar. 2007
[18] Xiaojing Yuan; Zhengyu Yang; Zouridakis, G.; Mullani, N.; , "SVM-based Texture
Classification and Application to Early Melanoma Detection," Engineering in Medicine and
Biology Society, 2006. EMBS '06. 28th Annual International Conference of the IEEE , vol.,
no., pp.4775-4778, Aug. 30 2006-Sept. 3 2006 .doi: 10.1109/IEMBS.2006.260056
[19] Steven McDonagh, Robert B. Fisher and Jonathan Rees,"Using 3D information for
classification of non-melanoma skin lesions",Available Online
homepages.inf.ed.ac.uk/rbf/PAPERS/mcdonagh.pdf
[20] M. EmreCelebi, Hassan A. Kingravia, Y. Alp Aslandogan, William V.
Stoecker,"Detection of Blue-White Veil Areas in Dermoscopy Images Using Machine
Learning Techniques",Proc. of SPIE Vol. 6144.doi: 10.1117/12.655779
[21] M. F. Healsmith, J. F. Bourke, J. E. Osborne, and R. A. C. Graham- Brown, “An
evaluation of the revised seven-point checklist for the early diagnosis of cutaneous malignant
melanoma,” Br. J. Dermatol., vol. 130, pp. 48–50, 1994
[22] R. B. Mallet, M. E. Fallowfield, M. G. Cook, W. N. Landells, C. A. Holden, and R. A.
Marsden, “Are pigmented lesion clinics worthwhile?” Br. J. Dermatol., vol. 129, pp. 689–
693, 1993
[23] Dangare, Chaitrali S. and Apte, Sulabha S., A Data Mining Approach for Prediction of
Heart Disease Using Neural Networks (November 14, 2012). International Journal of
Computer Engineering and Technology (IJCET), Volume 3, Issue 3, October-December 2012.
Available at SSRN: http://ssrn.com/abstract=2175569
[24] S. ShenbagaEzhil and Dr. C. Vijayalakshmi. "PREDICTION OF COLON-RECTUM
CANCER SURVIVABILITY USING ARTIFICIAL NEURAL NETWORK",International
Journal of Computer Engineering and Technology (IJCET),Volume 3, Issue 1, January- June
(2012), pp. 163-168
[25] Marghoob AA, Braun RP, Kopf AW. Atlas of dermoscopy. London: Taylor & Francis;
2005
[26] Ru-Song Meng, Xiao Meng, Feng-Ying Xie and Zhi-GuoJiang,"Early diagnosis for
cutaneous malignant melanoma based on the intellectualized classification and recognition
for images of melanocytic tumour by dermoscopy",Journal of Biomedical Graphics and
Computing, December 2012, Vol. 2, No. 2. pp 37-47
[27] S. Aruna, L.V. Nandakishore and Dr S.P. Rajagopalan, “A Novel Lns Semi Supervised
Learning Algorithm For Detecting Breast Cancer” International journal of Computer
Engineering & Technology (IJCET), Volume3, Issue1, 2012, pp. 44 - 53, Published by
IAEME
[28] S. Shenbaga Ezhil and Dr. C. Vijayalakshmi, “Prediction of Colon-Rectum Cancer
Survivability Using Artificial Neural Network” International journal of Computer
Engineering & Technology (IJCET), Volume3, Issue1, 2012, pp. 163 - 168, Published by
IAEME
33