International Journal of Trend in Scientific Research and Development (IJTSRD)

Volume: 3 | Issue: 4 | May-Jun 2019 Available Online: www.ijtsrd.com e-ISSN: 2456 - 6470

Melanoma Skin Cancer Detection using

Image Processing and Machine Learning
Vijayalakshmi M M
Assistant Professor, Department of Information Science & Engineering GSSSIETW, Mysuru, Karnataka, India

How to cite this paper: Vijayalakshmi M ABSTRACT

M "Melanoma Skin Cancer Detection Dermatological Diseases are one of the biggest medical issues in 21st century
using Image Processing and Machine due to its highly complex and expensive diagnosis with difficulties and
Learning" Published in International subjectivity of human interpretation. In cases of fatal diseases like Melanoma
Journal of Trend in Scientific Research diagnosis in early stages play a vital role in determining the probability of getting
and Development cured. We believe that the application of automated methods will help in early
(ijtsrd), ISSN: 2456- diagnosis especially with the set of images with variety of diagnosis. Hence, in
6470, Volume-3 | this article we present a completely automated system of dermatological disease
Issue-4, June 2019, recognition through lesion images, a machine intervention in contrast to
pp.780-784, URL: conventional medical personnel-based detection. Our model is designed into
https://www.ijtsrd.c three phases compromising of data collection and augmentation, designing
om/papers/ijtsrd23 IJTSRD23936 model and finally prediction. We have used multiple AI algorithms like
936.pdf Convolutional Neural Network and Support Vector Machine and amalgamated it
with image processing tools to form a better structure, leading to higher
Copyright © 2019 by author(s) and accuracy of 85%.
International Journal of Trend in
Scientific Research and Development Keywords: Dermatology, Image Processing, Machine Learning, Melanoma
Journal. This is an Open Access article
distributed under I. INTRODUCTION
the terms of the Skin is the outer most region of our body and it is likely to be exposed to the
Creative Commons environment which may get in contact with dust, Pollution, micro-organisms and
Attribution License (CC BY 4.0) also to UV radiations. These may be the reasons for any kind of Skin diseases and
(http://creativecommons.org/licenses/ also Skin related diseases are caused by instability in the genes this makes the
by/4.0) skin diseases more complex.
The human skin is composed of two major layers called Melanoma. Malignant Melanoma is one of the deadly and
epidermis and dermis. The top or the outer layer of the skin dangerous type cancers, even though it’s found that only 4%
which is called the epidermis composed of three types of of the population is affected with this, it holds for 75% of the
cells flat and scaly cells on the surface called SQUAMOUS death caused due to skin cancer. Melanoma can be cured if
cells, round cells called BASAL cells and MELANOCYTES, its identified or diagnosed in early stages and the treatment
cells that provide skin its color and protect against skin can be provided early, but if melanoma is identified in the
damage. As the diagnostic classification currently do not last stages, it is possible that Melanoma can spread across
represent the diversity of the disease, these are not sufficient deeper into skin and also can affect other parts of the body,
enough to make a correct prediction and also treatment to then it becomes very difficult to treat. Melanoma is caused
be provided for that disease. Adding to this cancer cells are due to presence of Melanocytes which are present with in
often diagnosed late and treated late, it is diagnosed when the body.
the cancer cells have mutated and spreads to the other
internal parts of the body. At this stage therapies or Exposure of skin to UV radiation is also one of the major
treatments are not very effective. Due to these kinds of reasons for the cause of Melanoma. Dermoscopy is a
issues skin cancer percentage is taken over by the heart technique, that is used to exam the structure of skin. An
related diseases as the most affected and it is the cause of observation-based detection technique can be used to detect
death among all ages in the world. The other reasons for Melanoma using Dermoscopy images. The accuracy of the
which the disease might have taken over to a very serious dermoscopy depends on the training of the dermatologist.
state can be because of people’s ignorance and also that The accuracy of Melanoma Detection can be 75%-85% even
people try using home remedies without knowing the though the experts in skin use dermoscopy as a method for
severity of the problem and also sometimes these may lead diagnosis. The diagnosis that is performed by the system will
to another kind of skin rashes or even increasing the severity help to increase the speed and accuracy of the diagnosis.
of the problem. Computer will be able to extract some information, like
asymmetry, color variation, texture features, these minute
Among all the types of skin diseases skin cancer is found to parameters may not be recognized by the human naked eyes.
be the deadliest kind of disease found in humans. This is There are 3 stages in an automated dermoscopy image
found most commonly among the fair skin. Skin cancer is analysis system, (a) pre-processing (b) Proper Segmentation,
found to be 2 types Malignant Melanoma and Non- (c) feature extraction and selection. The segmentation is the

@ IJTSRD | Unique Paper ID - IJTSRD23936 | Volume – 3 | Issue – 4 | May-Jun 2019 Page: 780
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
most important and also plays a key role as it affects the the clear enhancement for the further steps by increasing the
process of fore coming steps. Supervised segmentation efficiency of the model. Pre-processing includes the
seems to be easy to implement by considering the following:
parameters like shapes, sizes, and colors along with skin
Collection of the dataset
types and textures. This system-based analysis will reduce

Hair removal
the diagnosing time and increases the accuracy.
Dermatological Diseases, due to their high complexity,
Shading removal
variety and scarce expertise is one of the most difficult
Glare removal
terrains for quick, easy and accurate diagnosis especially in
developing and under-developed countries with low Dataset: The images were collected from the ISIC dataset;
healthcare budget. Also, it’s a common knowledge that the the ISIC dataset provide the collection of images for
early detection in cases on many diseases reduces the melanoma skin cancer. ISIC melanoma project was
chances of serious outcomes. The recent environmental undertaken to reduce the increasing deaths related to
factors have just acted as catalyst for these skin diseases. melanoma and efficiency of melanoma early detection. This
ISIC dataset contains approximately 23,000 images of which
The general stages of these diseases are as: STAGE 1- we have collected 1000-1500 images and trained and tested
diseases in situ, survival 99.9%, STAGE 2- diseases in high over these images.
risk level, survival 45-79%, STAGE 3-regional metastasis,
survival 24-30%, STAGE 4- distant metastasis-survival 7- Hair Removal: for the above collected images hair removal
19% method was applied this method was performed using
Hough transform, Hough transform is basically used to
II. RELATED WORKS
identify lines or elliptical or circular shapes. Performing hair
The authors [1] have tried to address the same problem
removal for the images that has hair within the tumor
using image analysis techniques. The work uses the
provides us an clear image of tumor which also helps us to
technique of noise removal and subsequent feature
make further more enhancements.
extraction. After the noise removal, the image is fed into
classifier for further feature extraction process and finally
the prediction of the disease. Most of the earlier publications Shading removal: The images that is taken from the dataset
focused on feature extraction and then subsequent disease contains shade around the region of the tumor this shade for
prediction was done. Papers [6,3] have used Artificial Neural few images is dark and for few is light, removal of the shade
Network for dealing with this complex problem while papers in the region of tumor also provides us an clear vision of the
[2,4,5] have used machine learning algorithms for the task. tumor which is also helpful in the further enhancements. We
Computer vision techniques have played a major role in have used the MATLAB filters to remove the shade for
many previous literatures. As is evident, the publishers have images in the dataset.
utilized the image processing techniques to accomplish the
pre processing task. In the similar way we also try to Glare Removal: sometime the images are captured from
implement the computer vision techniques, but out camera the images will contain glare this glare is not visible
implementation mainly focuses for dataset augmentation. to the naked eyes, we remove this glare using the MATLAB
filter, this minute noise sometimes may affect the accuracy at
III. Methodology the end.
Our model is designed in 3 phases as follows:
A. Phase1 – the first model involves collection of dataset, V. Architecture
the images are collected from ISIC dataset (International
Skin Imaging Collaboration) Phase 1 also involves the
pre-processing of the images where hair removal, glare
removal and shading removal are done
B. Removal of these parameters helps us to identify the
texture, color, size and shape like parameters in an
efficient way.
C. Phase2- this phase consists of the segmentation and
feature extraction, segmentation is explored via three
methods a. Otsu segmentation method b. Modified Otsu
segmentation method c. water shed segmentation
method. Feature are extracted for color, shape, size and
texture.
D. Phase 3- this is the most important phase of our model,
this phase involves designing of the model and training.
Our model was trained for Back Propagation Algorithm
(Neural Networks), SVM (Support Vector Machine), and
CNN (Convolutional Neural Networks) on the dataset VI. Designing The Model
that was collected in the phase1, the model after training In our model we have used 3 different methods i.e. Neural
was tested for the accurate output. Networks, Support Vector Machine and Convolutional Neural
Networks to find the efficient detection and classification of
IV. COMPONENTS OF METHODOLOGY: the melanoma skin cancer into Malignant and benign skin
PRE-PROCESSING: cancers. The data that is pre-processed is followed by
The pre-processing of images is an important task or activity segmentation and feature extraction these extracted feature
which helps in saving time for training as well as provides images are then passed into Neural Networks and Support

@ IJTSRD | Unique Paper ID - IJTSRD23936 | Volume – 3 | Issue – 4 | May-Jun 2019 Page: 781
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
Vector Machine to classify the images into malignant and We need to reach the ‘Global Loss Minimum’. This is nothing
benign and to predict the exact accuracy. but Backpropagation.

A. Neural Networks B. Support Vector Machine (SVM)

In the neural Networks we have used the Back Propagation SVM (Support Vector Machine) is a supervised machine
Algorithm. The Back Propagation is a supervised learning learning algorithm which is mainly used to classify data into
algorithm, for training the multi-layer perceptron’s. while different classes. Unlike most algorithms, SVM makes use of
designing the neural networks we initialize the weights with a hyperplane which acts like a decision boundary between
some random values as we do not know what exactly the the various classes. SVM can be used to generate multiple
weight can be, so we first give some random weight if the separating hyperplanes such that the data is divided into
model provides an error with large values. so, we need to segments and each segment contains only one kind of data.
need to change the values to somehow minimize the error Features of SVM are as follows:
value. To generalize this, we can just say 1. SVM is a supervised learning algorithm. This means that

Calculate the error – How far is your model output SVM trains on a set of labelled data. SVM studies the
from the actual output labelled training data and then classifies any new input

Minimum Error – Check whether the error is data depending on what it learned in the training phase.
minimized or not. 2. A main advantage of SVM is that it can be used for both

Update the parameters – If the error is huge then, classification and regression problems. Though SVM is
update the parameters (weights and biases). After that mainly known for classification, the SVR (Support Vector
again check the error. Repeat the process until the error Regressor) is used for regression problems.
becomes minimum. 3. SVM can be used for classifying non-linear data by using

Model is ready to make a prediction – Once the error the kernel trick. The kernel trick means transforming
becomes minimum, you can feed some inputs to your data into another dimension that has a clear dividing
model and it will produce the output. margin between classes of data. After which you can
easily draw a hyperplane between the various classes of
data.

What is support vectors in SVM? we start of by drawing a

random hyperplane and then we check the distance between
the hyperplane and the closest data points from each class.
These closest data points to the hyperplane are known as
support vectors. And that’s where the name comes from,
support vector machine.

In this project we have used SVM to classify the malignant

and benign skin cancer images, this done by passing the
The Backpropagation algorithm looks for the minimum value segmented and feature extracted images into SVM where
of the error function in weight space using a technique called SVM write the hyperplane and groups all the near by similar
the delta rule or gradient descent. features into different classes.

we are trying to get the value of weight such that the error
becomes minimum. Basically, we need to figure out whether
we need to increase or decrease the weight value. Once we
know that, we keep on updating the weight value in that
direction until error becomes minimum. You might reach a
point, where if you further update the weight, the error will
increase. At that time, you need to stop, and that is your final
weight value.

Consider the graph below:

The performance of the SVM classifier was very accurate for

even a small data set and its performance was compared to
other classification algorithms like CNN and Back
Propagation Algorithm.

C. Convolution Neural Network

CNNs are neural networks with a specific architecture that
have been shown to be very powerful in areas such as image
recognition and classification. CNNs have been

@ IJTSRD | Unique Paper ID - IJTSRD23936 | Volume – 3 | Issue – 4 | May-Jun 2019 Page: 782
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
demonstrated to identify faces, objects, and traffic signs
better than humans and therefore can be found in robots and
self-driving cars.

CNNs are a supervised learning method and are therefore

trained using data labeled with the respective classes.
Essentially, CNNs learn the relationship between the input
objects and the class labels and comprise two components:
the hidden layers in which the features are extracted and, at
the end of the processing, the fully connected layers that are
used for the actual classification task. Unlike regular neural
networks, the hidden layers of a CNN have a specific
architecture. In regular neural networks, each layer is
formed by a set of neurons and one neuron of a layer is
connected to each neuron of the preceding layer. The
architecture of hidden layers in a CNN is slightly different.
The neurons in a layer are not connected to all neurons of
the preceding layer; rather, they are connected to only a
small number of neurons. This restriction to local
connections and additional pooling layers summarizing local
neuron outputs into one value results in translation-
invariant features. This results in a simpler training
procedure and a lower model complexity

VII. CONCLUTION VIII. REFERENCES

The aim of this project is to determine the accurate [1] abrham debasu mengistu , dagnachew melesew
prediction of skin cancer and also to classify the skin cancer alemayehu “computer vision for skin cancer diagnosis
as malignant or non-malignant melanoma. To do so, some and recognition using rbf and som “ international
pre-processing steps were carried out which followed Hair journal of image processing (ijip), volume (9) : issue
removal, shadow removal, glare removal and also (6) 2015.
segmentation. SVM and Deep Neural networks will be used
[2] s.s. Mane1, s.v. Shinde “different techniques for skin
to classify. classifier will be trained to learn the features and
cancer detection using dermoscopy images” ,
finally used to classify. The novelty of the present
international journal of computer sciences and
methodology is that it should do the detection in very quick
engineering vol.5(12), dec 2017, e-issn: 2347-2693.
time hence aiding the technicians to perfect their diagnostic
skills. The dataset used is from the available ISIC [3] poornima m s, dr. Shailaja k “detection of skin cancer
(International Skin Image Collaboration) dataset, hence any using svm” , international research journal of
dataset can be used to find the efficiency. engineering and technology (irjet) volume: 04 issue: 07
| july -2017.
[4] yuexiang li and linlin shen “skin lesion analysis
towards melanoma detection using deep learning
network”, arxiv:1904.073653v2 [cs.cv] 20 aug 2018
[5] muhammad imran razzak, saeeda naz and ahmad zaib “
deep learning for medical image processing: overview,
challenges and future” arxiv:1852.3865v2 [cs.cv] 20
july 2018
[6] veronika cheplygina, marleen de bruijne, josien p. W.
Pluim, “ not-so-supervised: a survey of semi-
supervised, multi-instance, and transfer Learning in
medical image analysis” arxiv:1804.06353v2 [cs.cv] 14
sep 2018
[7] salome kazeminia, christoph baur, arjan kuijper, bram
van Ginneken, nassir navab, shadi albarqouni, anirban
mukhopadhyay “gans for medical image analysis “,
arxiv:1809.06222v2 [cs.cv] 21 dec 2018
[8] andreas maier, christopher syben, tobias lasser,
christian riess “a gentle introduction to deep learning
in medical image processing”, arxiv:1810.05401v2
[cs.cv] 21 dec 2018
[9] danilo barros mendes , nilton correia da silva “skin
lesions classification using convolutional Neural
networks in clinical images”, arxiv:1812.02316v1
[cs.cv] 6 dec 2018

@ IJTSRD | Unique Paper ID - IJTSRD23936 | Volume – 3 | Issue – 4 | May-Jun 2019 Page: 783
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
[10] wasan kadhim saa'd “ method for detection and [16] lakshay bajaj, himanshu kumar, yasha hasija,”
diagnosis of the Area of skin disease based on color by automated system for prediction of skin disease using
Wavelet transform and a rtificial neural Network” al- image processing and machine learning” international
qadisiya journal for engineering sciences vol. 2 no.4 journal of computer applications (0975 – 8887) volume
year 2009 180 – no.19, february 2018
[11] li-sheng wei , quan gan, and tao ji , “skin disease [17] ritesh maurya, surya kant singh, ashish k. Maurya, ajeet
recognition method based on image color and Texture kumar,” glcm and multi class support vector machine
features” hindawi computational and mathematical based automated skin cancer classification” ieee
methods in medicine volume 2018, article id 8145713,
[18] prashant b. Yadav, mrs. S.s. Patil “ recognition of
10 pages
dermatological disease area for identification of
[12] rahat yasir, md. Shariful islam nibir, and nova ahmed “ disease” ijsdr may 2016 volume 1, issue 5
a skin disease detection system for financially unstable
[19] nikita raut, aayush shah, shail vira, harmit sampat, “ a
people in developing countries” global science and
study on different techniques for skin cancer
technology journal vol. 3. No. 1. March 2015 issue. Pp. 77
detection”, international research journal of
– 93
engineering and technology (irjet), volume: 05 issue:
[13] t.yamunarani, “analysis of skin cancer using abcd 09 | sep 2018
technique” , international research journal of
[20] m.yuvaraju, d.divya, a.poornima, “segmentation of skin
engineering and technology (irjet) volume: 05 issue: 04
lesion from digital images using morphological filter”,
| apr-2018
international research journal of engineering and
[14] rahat yasir,, md. Ashiqur rahman, and nova ahmed, technology (irjet) volume: 03 issue: 05 | may-2016
“dermatological disease detection using image
[21] yuexiang liid and linlin shen, “skin lesion analysis
Processing and artificial neural network”,
toward melanoma detection using deep learning
arxiv:1012.2436v1 [cs.cv] 16 dec 2018
network” sensors mdpi 11 february 2018.
[15] m. Shamsul arifini, m. Golam kibria, adnan firoze, m.
[22] mrs. S kalaiarasi, harsh kumar, sourav patra,
Ashraful amini, hong yan, “dermatological disease
“dermatological disease detection using image
diagnosis using color-skin images”, proceedings of the
processing and neural networks”, s.kalaiarasi et al,
2012 international conference on machine learning
international journal of computer science and mobile
and cybernetics, xian, 15-17 july, 2012
applications, vol.6 issue. , pg. 109-118 ,4 april- 2018.

@ IJTSRD | Unique Paper ID - IJTSRD23936 | Volume – 3 | Issue – 4 | May-Jun 2019 Page: 784