Multimedia Tools and Applications

https://doi.org/10.1007/s11042-022-13081-x

Automated deep learning approach for classification

of malignant melanoma and benign skin lesions

Wessam Salma 1 & Ahmed S. Eltrass 2

Received: 14 December 2020 / Revised: 9 March 2021 / Accepted: 4 April 2022

# The Author(s) 2022

Abstract
Skin cancer becomes a significant health problem worldwide with an increasing incidence
over the past decades. Due to the fine-grained differences in the appearance of skin
lesions, it is very challenging to develop an automated system for benign-malignant
classification through images. This paper proposes a novel automated Computer Aided
Diagnosis (CAD) system for skin lesion classification with high classification perfor-
mance using accuracy low computational complexity. A pre-processing step based on
morphological filtering is employed for hair removal and artifacts removal. Skin lesions
are segmented automatically using Grab-cut with minimal human interaction in HSV
color space. Image processing techniques are investigated for an automatic implementa-
tion of the ABCD (asymmetry, border irregularity, color and dermoscopic patterns) rule
to separate malignant melanoma from benign lesions. To classify skin lesions into benign
or malignant, different pretrained convolutional neural networks (CNNs), including
VGG-16, ResNet50, ResNetX, InceptionV3, and MobileNet are examined. The average
5-fold cross validation results show that ResNet50 architecture combined with Support
Vector Machine (SVM) achieve the best performance. The results also show the effec-
tiveness of data augmentation in both training and testing with achieving better perfor-
mance than obtaining new images. The proposed diagnosis framework is applied to real
clinical skin lesions, and the experimental results reveal the superior performance of the
proposed framework over other recent techniques in terms of area under the ROC curve
99.52%, accuracy 99.87%, sensitivity 98.87%, precision 98.77%, F1-score 97.83%, and

* Ahmed S. Eltrass
ahmed.eltrass@alexu.edu.eg

Wessam Salma
Wessam.salama@pua.edu.eg

1
Basic Science Department, Faculty of Engineering, Pharos University, Alexandria, Egypt
2
Electrical Engineering Department, Faculty of Engineering, Alexandria University, Alexandria,
Egypt
 Multimedia Tools and Applications

consumed time 3.2 s. This reveals that the proposed framework can be utilized to help
medical practitioners in classifying different skin lesions.

1 Introduction

Skin cancer has been increasing among men and women worldwide for many decades [32].
There were approximately 76,250 new cases of melanoma and approximately 8,790 new
melanoma-related deaths during 2012 in the United States [51]. In Brazil, it is estimated that
for the biennium of 2018–2019, there were 165,580 new cases of non-melanoma skin cancer
[34]. Skin cancer spreading results from many factors such as long longevity of the population,
people being exposed to the sun, and early detection of skin cancer. Dermoscopy, a noninva-
sive skin imaging technique, is one of the most effective ways for early identification of skin
cancer. The appearance of skin lesions in dermoscopic images may change significantly
according to the skin condition. In addition, different artifact sources, including hair, skin
texture, and air bubbles may lead to misidentifying the boundary between the skin lesions and
the surrounding healthy skin.
Despite the effectiveness of dermoscopic diagnosis for skin cancer, it is very difficult for
expert dermatologists to provide an accurate classification of malignant melanoma and benign
skin lesions for large number of dermoscopic images. Hence, it is very necessary to build up a
non-invasive Computer Aided Diagnosis (CAD) system for skin lesion classification. The
CAD system generally comprises of four main steps: image pre-processing, segmentation,
feature extraction, and classification. Note that each step significantly affects the classification
performance of the whole CAD system [50]. Therefore, efficient algorithms should be
employed in each step to achieve high diagnosis performance.
Various studies examined distinct machine learning methods for diagnosis of different
types of cancer [7, 11, 33, 45]. The majority of these studies used classifiers trained on a set of
hand-crafted features captured from the images. Most of machine learning techniques require
high computational time for accurate diagnosis and their performance depends on the selected
features that characterize the cancerous region. Deep learning techniques and Convolutional
Neural Networks (CNNs) have become important for automated diagnosis of different types of
cancer [3]. Deep learning has achieved impressive results in image classification applications.
In image classification tasks, transfer learning [29] and data augmentation [38] are employed
to overcome the lack of data and to reduce the computational and memory requirements.
Transfer learning is an effective tool for classification problems with few available datasets
like medical imaging applications. Instead of training a CNN from scratch, which would
require a huge amount of data and a high computational cost, it is computationally efficient to
employ a pre-trained CNN architecture and just fine-tune its performance to accelerate the
operation. Various pre-trained CNNs, including AlexNet, Inception, ResNet, and DenseNet [8,
37] were trained on the ImageNet Large Scale Visual Recognition Challenge (ILSVRC)
dataset. Some studies employed distinct pre-trained CNNs and reported reasonable diagnosis
performance for skin cancer [2, 13, 26, 30, 31, 36].
Data augmentation is a technique for enhancing the size of the input data by developing
new data from the original data [38]. Image augmentation techniques can be employed to
defeat the lack of skin cancer datasets. There are many data augmentation strategies, including
rotation, scaling, random cropping, and color modifications. Data augmentation is broadly
employed with pre-trained CNN architectures. In [30], the performance of skin lesion classi-
fication was investigated with 13 data augmentation techniques trained on three CNNs
Multimedia Tools and Applications

(Inception-v4, ResNet, and DenseNet), and the results revealed the positive impact of using
data augmentation in both training and testing phases. An automatic skin lesion classification
system based on the Alex-Net CNN architecture was proposed by [13]. The architecture
weights were fine-tuned, and the datasets are augmented by fixed and random rotation angles,
yielding average accuracy of 95.91%.
In [26], a deep learning approach based on neural network ensemble was employed to
classify skin lesions using dermoscopic images. The melanoma classification system of [31]
was developed for extracting asymmetry, border irregularity, color variation, diameter, and
texture of the skin lesion and classifying malignant melanoma and benign skin lesions using
Support Vector Machine (SVM) classifier. In [2], a deep CNN prediction model based on new
regularizer technique was employed for skin lesion classification with an accuracy of 97.49%.
In the work of [36], a Gabor wavelet-based CNN model for skin lesion diagnosis was
developed by employing Gabor filters on skin images to obtain detailed characteristics of
the skin lesions and then modeling these details using deep CNN models. Decision fusion
based on the sum rule was used for classifying the skin lesions with an accuracy of 83%.
The majority of exiting skin lesion diagnosis systems report reasonable classification results
for discriminating malignant melanoma from benign lesions. However, it is still very difficult
to develop an efficient automated framework that can provide very accurate diagnosis results,
while running in real-time using low hardware specifications. The aim of this study is to
propose an accurate diagnosis system for skin lesion classification in dermoscopic images. The
proposed approach is based on employing efficient image processing techniques for pre-
processing, data augmentation, and image segmentation along with investigating different
pre-trained CNN architectures for skin lesion classification. The proposed framework is
examined on the most common dermoscopic image databases, the first one is the International
Skin Imaging Collaboration (ISIC) 2017 [15], and the second one is the MNIST: HAM10000
[17].
The main contributions of this work can be summarized as follows: (1) A new real-
time automated CAD system is proposed for skin lesion classification with high
performance using low computational complexity. (2) The proposed framework is based
on new pre-processing techniques for hair removal and lesion identification, Grab-cut in
HSV color space for lesion segmentation, and automatic detection of the ABCD
features for discriminating malignant melanoma from benign lesions. (3) Different
pretrained CNNs, including VGG-16, ResNet50, ResNetX, InceptionV3, and MobileNet
are examined for skin lesion classification. (4) The impact of applying data augmenta-
tion to all pretrained CNN models under investigation (VGG-16, ResNet50, ResNetX,
InceptionV3, and MobileNet) is examined using several evaluation metrics, including
area under the ROC curve, accuracy, sensitivity, precision, F1-score, and computational
time using 5-fold cross-validation. The results revealed the positive influence of utiliz-
ing data augmentation in improving the diagnosis performance. Extensive experiments
have been performed to evaluate the performance of the proposed framework, and
experimental results reveal the superior performance of the proposed approach over
other exiting state-of-the-art diagnosis systems in terms of the classification rate, the
sensitivity, the specificity, and the computational efficiency. This demonstrates that the
proposed framework represents an efficient tool to assist dermatologists in real-time
skin lesion classification.
The remaining parts of this paper are structured as follows. Section 2 presents the
techniques used in pre-processing, segmentation, feature extraction, and classification.
 Multimedia Tools and Applications

Section 3 shows the experimental results of the proposed skin lesion classification system
using real dermoscopic images, followed by a comparison between the proposed framework
and other recent methods. In Section 4, the conclusions are presented.

2 Methodology

The proposed framework for skin lesion classification comprises of four main stages: image
pre-processing, segmentation, feature extraction, and classification. The structure of the
proposed approach is shown in Fig. 1.

2.1 Image pre-processing

The pre-processing step is used to eliminate all artifact sources such as bubbles and hair from
skin images. This step is very essential before any segmentation or feature extraction to
provide accurate diagnosis. A new approach based on morphological filtering is investigated
to remove the hair from the skin images [35]. This technique is divided into two mainly steps.
First, the color images are converted into grayscale versions using the weighted process of
[47], where the grayscale image is obtained from the RGB color space by the relation:
grayscale ¼ 0:3R þ 0:59G þ 0:11B. Second, the hair contour of the grayscale images
is identified using morphological black-hat transformation [52]. Finally, the inpainting
function is used to create a mask based on the Fast-Marching Method (FMM) [42]. In
the current work, when the pixel value is lower than the threshold, it will be set to 0,
otherwise it will be set to 1.

Fig. 1 The structure of the proposed skin cancer CAD system

Multimedia Tools and Applications

After the pre-processing step, each pre-processed skin image is augmented into four images
by rotating the input image into four directions of 0°, 90°, 180°, and 270°. Data augmentation
is investigated in the current work to increase the data size, generate new data from the original
input data, and overcome the lack of the tagged images.

2.2 Grab-cut skin lesion segmentation

After the pre-processing step, the images are segmented automatically using Grab-cut seg-
mentation technique [22]. In this technique, the boundary regions are identified, i.e., region
between the lesion, background, and the difference within the lesion. All skin images are
converted into HSV color space to distinguish image intensity from color information which is
a significant step for the adaptive histogram equalization process as shown in Fig. 2. Using the
Gaussian Mixture Model (GMM) [54], the foreground region is extracted from the Grab-cut
mask resulted from the adaptive histogram equalization process. Any region beyond this
rectangle will be taken as background. The segmentation results are quantitively evaluated
by determining the overlap with the ground truth. Generally, Jaccard Coefficient (JC) is the
most frequently used evaluation metric in image segmentation analysis. The JC measures
pixels similarity between the ground truth and the segmented image to examine which pixels
are matched and which are unmatched. It is defined as the intersection over union between
index segmentation results and ground truth [16]. Its value lies between 0 and 1, with 1
showing perfect overlap and 0 indicating no overlap. The segmented images with their JC
values are shown in Table 1.

Fig. 2 Conversion from RGB into HSV color space

 Multimedia Tools and Applications

Table 1 Segmentation results using Grab-cut technique

Original image Ground Truth M ask Segmented Image JC

0.89

0.82

2.3 Feature extraction

To distinguish malignant melanoma from benign skin lesions, the main features of skin lesions
should be extracted. The ABCD rule of dermoscopy [28] is investigated in the current study
for skin lesion identification. The ABCD features are extracted from the intrinsic physical
attributes of skin lesions, including asymmetry, border irregularity, color variation, and
dermoscopic patterns. The ABCD rule is used to characterize the geometrical and structural
lesion properties and to separate melanoma from benign lesions because of its effectiveness,
efficiency, and simplicity of performance and implementation. Table 2 shows the results of
ABCD rule of dermoscopy. The four main components of this method can be summarized as
follows [19]:
& Asymmetry A: Axes of asymmetry are obtained by drawing two orthogonal axes crossing
at the center of gravity of the lesion. For a given axis, the determination of whether a lesion
is symmetric or not is identified based on the symmetry along that axis for all measures,
namely color, brightness, and shape. The asymmetry score is zero if there is no asymmetry
in both axes. A score of one is given if there is asymmetry by one axis, and a score of two
is given in the case of asymmetry by both axes.
& Border B: The lesion is partitioned into eight parts (slices). A border score is assigned
according to the existence of sharp or gradual cut-off at the periphery of the skin lesions. A
score of zero is given for regular periphery, while a score of one is given for irregular
boundary. The border score varies between zero and eight according to the number of
slices with irregular periphery. Vector product and inflexion point descriptors are utilized
to extract the number of peaks, valleys, and straight lines at lesion edges corresponding to
clear irregular and small irregular peripheries, respectively [10].
& Color C: It is one of the most important discriminators of malignant melanoma from
benign lesions. Lesions with at least one of six colors, namely white, red, black, light
brown, dark brown, and blue-gray, are probably melanomas. The existence of one
suspicious color is confirmed when the number of pixels concerning that color surpasses
by 5% of the total number of pixels of the lesion [43]. In a lesion, more than one suspicious
Multimedia Tools and Applications

Table 2 Feature extraction using ABCD rule of dermoscopy

Asymmetry

Border structure

Color variation

Dermoscopic
structures

color may occur, and the color score is increased by one for each existing color. The color
score ranges from one to six.
& Dermoscopic structures D: The existence of five distinct structures, namely network,
structureless areas, branched streaks, dots, and globules, within the lesion is investigated.
The existence of each structure within the lesion increments the score by one. The
minimum structure score is zero and the maximum structure score is five.

2.4 Classification with CNN

In this step, the resulted ABCD dermoscopy images are fed into a pretrained deep CNN
architecture. Different pretrained CNN architectures, including VGG-16 [39], ResNet50 [18],
ResNetX [9], InceptionV3 [41], and MobileNet [14] are investigated for skin lesion classifi-
cation. Note that designing a new deep CNN network from scratch is hard because (i) it needs
a huge amount of labeled training data and high experience to guarantee efficient convergence;
(ii) the process is time-consuming and tedious; (iii) it may suffer from overfitting problems.
Instead of training a CNN from scratch, it is computational efficient to employ a pre-trained
CNN model and just fine-tune its performance to accelerate the operation. All pre-trained
CNNs considered in the current study were trained on a huge dataset, namely ILSVRC. They
were trained on approximately 1.2 million training images with another 50,000 images for
validation and 100,000 images for testing. In this work, we use the weights pretrained on
 Multimedia Tools and Applications

ImageNet to fine-tune the model. This makes the pretrained network can learn new task faster
and easier than the network which is trained from scratch using randomly initialized weight.
For all pretrained CNN architectures under investigation, classification results are exam-
ined, and the results reveal that VGG-16 and ResNet50 architectures achieve the best
performance. In this work, a batch size of 40/16, learning rate of 103 /104 , and momentum
of 0.8/0.9 are, respectively, the best values for yielding the highest diagnosis performance of
ResNet50 and VGG-16 architectures. The detailed layers information of ResNet50 and VGG-
16 architectures are shown in Tables 3 and 4, respectively. The proposed framework is
enhanced by combining the linear kernel SVM classifier [4] with both VGG-16 and ResNet50
architectures and investigating their classification results.
To authenticate the diagnosis performance and avoid any bias, the k-fold cross-validation
technique is employed [40]. Usually, the value of k is aprioristically selected and fixed.
Generally, increasing the value of k from 2 to 10 leads to bias decrease, variance increase in
the test error rate estimation, and more computational burden [25]. Note that the computational
time tends to follow an increasing exponential distribution with the increase of k. Larger values
of k are usually employed to use a larger number of samples for training at the expense of error
estimation degradation of the classifier [12]. In the current study, the training data is sufficient
to use 5-fold cross validation which allows reducing the variance estimate of error prediction
and minimizing the computational cost. This is very critical for practical implementation of the
proposed CAD system in real-time applications. Also, obtaining a tight and rigorous estima-
tion of the classifier error allows validating, in a statistical sense, the reliability of the model
[12]. Moreover, when employing the 5-fold cross validation, larger fraction of samples can be
kept for the classification phase. For the data under investigation, the 5-fold cross validation
represents a good choice to compromise between the percentage of data used for training, and
the rigor of the estimated error.

Table 3 Summary of the CNN layers for the ResNet50 structure

Layer No. Type Size Kernel/Stride Activation

Input 224*224*3 ------ -------

1 Convolution 64*3*3 2 ReLU
2 Convolution 1*1*64 2 ReLU
Convolution 3*3*64
Convolution 1*1*256
3 Convolution 1*1*128 2 ReLU
Convolution 3*3*128
Convolution 1*1*512
4 Convolution 1*1*256 2 ReLU
Convolution 3*3*256
Convolution 1*1*1024
5 Convolution 1*1*512 2 ReLU
Convolution 3*3*512
Convolution 1*1*2048
Output 2 ---- SVM
Multimedia Tools and Applications

Table 4 Summary of the CNN layers for the VGG-16 structure

Layer No. Type Size Kernel size Activation

Input 224*224*3 ----- -----

1 Convolution 224*224*64 ------ ReLU
Convolution 224*224*64 ------
MaxPooling 112*112*64 2*2
2 Convolution 112*112*128 ----- ReLU
Convolution 112*112*128 -----
MaxPooling 56*56*128 2*2
3 Convolution 56*56*256 ----- ReLU
Convolution 56*56*256 ----
Convolution 56*56*256 ----
MaxPooling 28*28*256 2*2
4 Convolution 28*28*512 ---- ReLU
Convolution 28*28*512 ----
Convolution 28*28*512 ---
MaxPooling 14*14*512 2*2
5 Convolution 14*14*512 ---- ReLU
Convolution 14*14*512 ----
Convolution 14*14*512 ---
MaxPooling 14*14*512 2*2
6 FC 4096*1 ------ ReLU+Dropout (0.5)
7 FC 4096*1 ----- ReLU+Dropout(0.5)
8 FC 4096*1 ----- ReLU+Dropout(0.5)
Output 2 SVM

3 Results

To examine the usefulness of the proposed skin lesion classification approach, 580
images (310 benign and 270 malignant) from the ISIC 2017 database and 350 images
(215 benign and 135 malignant) from the MNIST: HAM10000 database are investi-
gated. The data under investigation covers distinct cases of abnormality, patient age,
and patient gender as shown in Table 5. In order to increase the total number of
images and allow more data for training and testing, each image is augmented into
four images of different directions of rotation, providing a total number of 2320
image from the ISIC 2017 database and 1400 image from the MNIST: HAM10000

Table 5 Distribution of images taken from the ISIC 2017 and MNIST: HAM10000 databases

ISIC2017 Database: 580 Images (270 Malignant and 310 Benign)

Patients Age (Years)
30–40 41–50 51–60 61–70 >70
90 120 210 100 60
Patients Gender (Male and Female)
Male 236
Female 344
HAM10000 Database: 350 Images (135 Malignant and 215 Benign cases)
Patients Age (Years)
30–40 41–50 51–60 61–70 >70
60 105 65 80 40
Patients Gender (Male and Female)
Male 193
Female 157
 Multimedia Tools and Applications

database. Evaluation of image classification depends mainly on computing four pa-

rameters: the number of true positives (TP), true negatives (TN), false negatives (FN),
and false positives (FP). The classification performance is identified in terms of
Accuracy ð ACC Þ, Sensitivity (Se) or recall, Precision (PrÞ, F1-score, area under the
ROC curve (AUC), and computational time. ACC is a metric of true predictions, Pr is
the fraction of detected malignant cases that agree with the ground truth, and Se is the
fraction of true malignant cases that are identified malignant. F1-score is the harmonic
mean of Pr and Se. AUC measures the entire two-dimensional area underneath the
entire ROC curve and it gives a total performance evaluation across all possible
classification thresholds. All these metrics are defined as follows:

ACC ¼ ðTP þ TN Þ=ðTP þ FP þ TN þ FNÞ ð1Þ

Pr ¼ TP=ðTP þ FPÞ ð2Þ

Se ¼ TP=ðTP þ FNÞ ð3Þ

F1 ¼ 2ðPr  SeÞ=ðPr þ SeÞ ð4Þ

All images under investigation (2320 augmented images from the ISIC 2017 and 1400
augmented images from the MNIST: HAM10000) are divided into 2 main partitions: 1120 for
training and validation and 280 for testing using the MNIST: HAM10000 database, and 1856
for training and validation, and 464 for testing using the ISIC 2017 database. To authenticate
the diagnosis performance and avoid any possible bias, the 5-fold cross-validation is utilized,
where the training and validation data (1856 from MNIST: HAM10000 database and 1856
from the ISIC 2017 database) are partitioned into five equal sized groups, each time one group
is retained as a validation set and the remaining four groups are retained as a training set. The
classification metrics are computed for each fold relevant to the test dataset for both datasets.
For all pretrained CNN architectures under investigation (VGG-16, ResNet50, ResNetX,
InceptionV3, and MobileNet), training and validation loss curves along with training and
validation accuracy curves are investigated, and the results reveal that VGG-16 and ResNet50
architectures achieve the best performance. For example, the training progress versus the
number of epochs for the ResNet50 architecture is shown in Fig. 3a. The ResNet50 network
has faster and better training advance than other CNN networks. Also, it achieves higher
performance than other CNN architectures in terms of the validation accuracy as shown in
Fig. 3a. The VGG-16 architecture achieves comparable performance with ResNet50 network.
Therefore, a critical comparison is made between VGG-16 and ResNet50 architectures for skin
lesion classification under various parameter regimes.
Table 6 shows the 5-fold benign-malignant classification results and standard deviations of
the proposed system using VGG-16 and ResNet50 architectures for the ISIC 2017 database
with data augmentation. The 5-fold average classification results ofAUC, ACC, Pr, Se, and F1-
score are calculated by averaging the 5-fold classification results (see Table 6). The average
Multimedia Tools and Applications

(a)

(b)

Fig. 3 Training and validation (a) accuracy and (b) loss versus number of epochs for the ResNet50 architecture

classification results for each metric were calculated for all CNN architectures under investi-
gation (VGG-16, ResNet50, ResNetX, InceptionV3, and MobileNet), showing the better
performance of ResNet50 architecture in terms of all metrics (ACC, AUC, Pr, Se, and F1-
score). Similar results are obtained for all CNN architectures using MNIST: HAM10000
database, and the results show the higher performance of ResNet50 architecture as shown in
Table 7.
We further enhance the proposed system by combining the kernel SVM classifier with all
CNN architectures and investigating their classification results. Table 7 shows the average 5-
fold benign-malignant classification results and standard deviations of VGG-16, ResNet50,
ResNetX, InceptionV3, and MobileNet architectures with and without SVM for both ISIC
2017 and MNIST: HAM10000 databases with and without data augmentation. It can be noted
that the results of either VGG-16 or ResNet50 architecture with SVM are better than the
standalone VGG-16 or ResNet50. Also, the performance of ResNet50 architecture combined
with SVM is superior to VGG-16 architecture with SVM in terms of average ACC(99.87%
versus 98.66%), AUC (99.52% versus 97.63%), Se (98.87% versus 95.49%), Pr (98.77%
 Multimedia Tools and Applications

Table 6 Average 5-fold classification results and standard deviation for ResNet50 and VGG-16 using 2320
images from the ISIC 2017 database

ResNet50 Architecture
Fold ACC AUC Se Pr F1-score
1 95.62% 93.42% 95.98% 93.94% 92.42%
2 95.78% 93.11% 94.88% 95.58% 94.98%
3 95.84% 92.44% 94.98% 95.68% 93.98%
4 95.75% 93.45% 93.99% 93.87% 92.99%
5 96.56% 92.47% 93.89% 94.99% 93.87%
Average 5-Fold Results 95.91% ± 0.26 92.97% ± 0.29 94.74% ± 0.31 94.81% ± 0.28 93.64% ± 0.27
VGG-16 Architecture
Fold ACC AUC Se Pr F1-score
1 97.75% 93.59% 94.84% 95.54% 94.98
2 96.45% 92.87% 92.98% 93.45% 93.95%
3 96.61% 92.89% 93.45% 91.81% 92.98%
4 96.77% 91.98% 94.74% 92.97% 91.97%
5 95.93% 92.87% 92.85% 93.89% 93.85%
Average 5-Fold Results 96.70% ± 0.36 92.84% ± 0.34 93.77% ± 0.31 93.53% ± 0.35 93.54% ± 0.30

Table 7 Average 5-fold classification results and standard deviations for VGG-16, ResNet50, ResNetX,
InceptionV3, and MobileNet architectures with and without SVM for both ISIC 2017 and MNIST:
HAM10000 databases with and without data augmentation

ISIC 2017 database with data augmentation

Model ACC AUC Se Pr F1-score
ResNet50+SVM 99.87% ± 0.21 99.52% ± 0.27 98.87% ± 0.25 98.77% ± 0.23 97.83% ± 0.21
ResNet50 95.91% ± 0.26 92.97% ± 0.29 94.74% ± 0.31 94.81% ± 0.28 93.64% ± 0.27
VGG16 +SVM 98.66% ± 0.27 97.63% ± 0.29 95.49% ± 0.31 96.79% ± 0.26 95.77% ± 0.25
VGG16 96.70% ± 0.36 92.84% ± 0.34 93.77% ± 0.31 93.53% ± 0.35 93.54% ± 0.30
ResNetX +SVM 93.65%± 0.21 93.82% ± 0.26 93.52% ± 0.24 93.64% ± 0.22 93.69% ± 0.23
ResNetX 92.68% ± 0.14 92.69% ± 0.17 92.78% ± 0.15 92.98% ± 0.11 92.79% ± 0.19
InceptionV3 +SVM 90.89% ± 0.18 90.78% ± 0.17 91.21% ± 0.13 90.89% ± 0.18 90.99% ± 0.16
InceptionV3 87.58% ± 0.32 87.34% ± 0.35 87.34% ± 0.31 88.02% ± 0.36 87.36% ± 0.36
MobileNet+SVM 85.68% ± 0.26 85.83% ± 0.29 86.02% ± 0.23 85.89% ± 0.28 86.57% ± 0.29
MobileNet 83.99% ± 0.38 84.98% ± 0.32 85.03% ± 0.31 84.68% ± 0.33 84.98% ± 0.35
ISIC 2017 database without data augmentation
ResNet50+SVM 96.98% ± 0.20 96.55% ± 0.21 96.96% ± 0.23 96.87% ± 0.21 95.98% ± 0.20
VGG16 +SVM 94.87% ± 0.28 93.96% ± 0.29 92.48% ± 0.30 92.93% ± 0.27 93.83% ± 0.28
ResNetX +SVM 91.48% ± 0.14 91.65% ± 0.18 91.69% ± 0.18 91.88% ± 0.15 91.98% ± 0.13
InceptionV3 +SVM 89.78% ± 0.11 89.65% ± 0.18 89.79% ± 0.16 88.99% ± 0.15 89.32% ± 0.19
MobileNet+SVM 84.32% ± 0.21 83.99% ± 0.22 84.65% ± 0.24 84.68% ± 0.26 84.83% ± 0.27
MNIST: HAM10000 database with data augmentation
ResNet50+SVM 98.79% ± 0.35 97.91% ± 0.33 97.78% ± 0.31 96.86% ± 0.29 97.45% ± 0.27
ResNet50 96.78% ± 0.37 96.43% ± 0.36 95.43% ± 0.33 93.86% ± 0.31 93.97% ± 0.29
VGG16 +SVM 94.74% ± 0.39 94.88% ± 0.35 96.14% ± 0.37 93.98% ± 0.32 92.45% ± 0.31
VGG16 92.76% ± 0.35 92.82% ± 0.31 92.69% ± 0.33 91.86% ± 0.29 92.97% ± 0.26
ResNetX +SVM 92.62% ± 0.15 92.32% ± 0.14 92.69% ± 0.11 92.65% ± 0.15 92.98% ± 0.15
ResNetX 93.21% ± 0.37 92.99% ± 0.36 93.21% ± 0.31 93.26% ± 0.38 93.12% ± 0.35
InceptionV3 +SVM 88.88% ± 0.15 87.99% ± 0.13 88.76% ± 0.11 88.68% ± 0.15 88.87% ± 0.17
InceptionV3 86.88% ± 0.37 86.99% ± 0.36 86.74% ± 0.33 86.69% ± 0.31 87.01% ± 0.35
MobileNet+SVM 84.84% ± 0.23 84.78% ± 0.21 84.89% ± 0.29 84.99% ± 0.23 84.87% ± 0.28
MobileNet 83.56% ± 0.15 83.69% ± 0.11 83.89% ± 0.13 83.63% ± 0.19 83.78% ± 0.12
MNIST: HAM10000 database without data augmentation
ResNet50+SVM 95.97% ± 0.32 94.98% ± 0.30 94.86% ± 0.30 95.83% ± 0.28 95.97% ± 0.27
VGG16 +SVM 94.89% ± 0.35 94.96% ± 0.33 93.87% ± 0.32 92.78% ± 0.31 92.68% ± 0.28
ResNetX +SVM 89.89% ± 0.14 90.01% ± 0.19 90.28% ± 0.18 89.99% ± 0.15 90.07% ± 0.13
InceptionV3 +SVM 86.87% ± 0.12 86.79% ± 0.19 86.96% ± 0.18 86.76% ± 0.15 86.86% ± 0.13
MobileNet+SVM 82.93% ± 0.23 82.77% ± 0.20 82.84% ± 0.26 82.65% ± 0.21 82.98% ± 0.24
Multimedia Tools and Applications

versus 96.79%), and F1-score (97.83% versus 95.77%) for ISIC 2017 database. The same
superior performance of ResNet50 architecture with SVM is achieved for the MNIST:
HAM10000 database. Note that the ResNet50 architecture with SVM provides smaller
standard deviation values than VGG-16 with or without SVM, showing its reliability and
consistency in skin lesion classification. In comparing the classification results with and
without data augmentation for both databases, results reveal that data augmentation allows
increasing the size of data used for training and testing, and hence enhancing the classification
results for all cases.

3.1 Performance analysis

The proposed ResNet50 CNN architecture combined with SVM for skin lesion classification is
compared with other recent systems [1, 2, 5, 6, 13, 19, 20, 23, 24, 30, 31, 36, 46, 48, 49, 53],
and the results are shown in Table 8. The average classification results using 5-fold cross-
validation are presented to validate the performance of the proposed system when comparing it
with other techniques. Table 8 reveals that we investigated larger number of lesion images than
most of other techniques under comparison. It can be noted that the proposed approach
outperforms other diagnosis systems in terms of ACC (99.87% for ISIC 2017 and 98.97%
for HAM10000) and AUC (99.52% for ISIC 2017 and 97.91% for HAM10000). The proposed
system achieves the highest Se in comparing with other techniques [1, 2, 13, 19, 20, 24, 31, 46,
48, 49, 53], revealing its effectiveness in classifying patients with skin cancer. Note that
sensitivity is a more important evaluation metric than other metrics for dermatologists and
clinicians because low sensitivity or high rate of false negative (malignant identified as benign)
may result in death. In this study, very few malignant lesions are misclassified, and conse-
quently the proposed approach has superior sensitivity results. Table 8 demonstrates the
superior diagnosis results of the proposed CAD system over recent state-of-the-art methods.
Different pretrained CNN architectures, including VGG-16, ResNet50, ResNetX,
InceptionV3, and MobileNet were investigated for skin cancer diagnosis, and the results show
that ResNet50 architecture achieves the best performance. A detailed comparison between all
these CNN architectures is made, and the results demonstrate the better performance of
ResNet50 architecture with low computational requirements. It is well-known that there are
various pretrained CNN structures [27], and we have already examined most of them. More
deeper CNN models mean more computational burden. This is not practical for real-time
diagnosis systems. Also, the use of more deep layers increases the number of free parameters,
which may lead to over-fitting issues and performance declination. In the current study, the
data variability is smaller than other image classification applications, and hence deeper CNNs
are not computationally effective for real-time diagnosis. The CNN models selected for this
study represent proper trade-off between speed and accuracy.
The overall average computational time of the proposed ResNet50 architecture combined
with SVM for skin lesion classification is around 3.23 s on a Kaggle Notebook GPU cloud (2
CPU cores and 13 GB RAM) using Python software. In comparing with other studies under
comparison, most of them did not report any computational time analysis. In [53], they
reported the computational time using NVIDIA GTX Titan XP GPU which was used for
training the model in round 30 h with an average test time of 0.02 s per patch. Despite the low
computational time of [53] in comparing to the proposed approach, it requires high hardware
requirements and provides low classification performance (see Table 8). The proposed ap-
proach achieves high diagnosis performance, while achieving reasonable computational
 Multimedia Tools and Applications

Table 8 Comparison between the proposed diagnosis system and several recent classification methods based on
different image processing techniques and CNN architectures using different datasets

Reference Number Databases

of lesion AUC ACC % Se %
images %

The proposed diagnosis system 2320 ISIC 2017 99.52 99.87 98.87
1400 HAM10000 97.91 98.79 97.78
Pretrained CNNs (Inception-v4, ResNet, and DenseNet) 2000 ISIC 2017 88.20 ------- -------
with different data augmentation techniques [30],
2018.
Pretrained Alex-Net architecture with data augmentation 2000 ISIC 2017 ------- 95.91 88.47
[13], 2019.
ABCD features, feature selection with principal 282 Private 88.10 82.2 77.02
component analysis (PCA), SVM classifier [31],
2014.
Deep CNN prediction model based on new regularizer 1760 ISIC archive 98.00 97.49 94.30
technique [2], 2019.
Gabor wavelet-based deep CNN [36], 2019. 2000 ISIC 2017 96.00 83.00 -------
Automatic ABCD scoring of skin lesions, [19], 2016. 200 Private ------- 94.00 91.20
Fully convolutional residual network [49], 2016. 1250 ISBI 2016 ------- 94.40 91.10
Optimal color and texture features, majority-voting 350 Atlas of 94.00 93.00 94.00
multi-class SVM [1], 2016. Dermoscopy
CD
Pre-processing with cytological properties, CNN with 1760 ISIC archive 84.70 86.60 80.90
5-fold cross validation [48], 2016.
Deep CNN with data augmentation [6], 2016. 211 ISIC 2017 ------- 88.00 -------
CNN model based on attention residual learning [53], 2750 ISIC 2017 87.50 85.00 65.80
2019.
Ant colony-based segmentation, geometric and texture 172 DermIS and ------- 93.60 -------
features, ANN classifier [5], 2017. DermQuest
Probabilistic distribution segmentation, entropy-based 200 PH2, ------- 97.50 96.67
method feature selection, multi-class SVM [20], 290 ISIC archive ------- 97.75 96.60
2018. 2750 ISBI 2016 96.00 93.2 93.00
&2017
Deep CNN with data augmentation [46], 2017. 900 ISBI 2016 ------- 83.60 76.00
Deep CNN combined with hand-crafted features and 1279 ISIC archive 78.00 82.60 53.33
local binary patterns [24], 2016.
Optimized deep features from well-established CNNs 2037 ISIC 2017 83.83 ------- -------
and from different abstraction levels [23], 2019.

efficiency using low hardware specifications. In comparing with the deep learning model of
[21] which achieved high ACC of 94.5%, the computational efficiency of the proposed
framework outperforms it in terms of testing time (3.23 s versus 6.9 s). This elucidates the
usefulness of the proposed framework in improving the skin lesion classification performance
through real-time systems.
Despite the high performance of the proposed skin lesion classification system, there are
few limitations that can be investigated in the future. First, the number of pre-trained CNN
architectures under investigation can be increased to incorporate more recent and advanced
pre-trained models. Second, the process of extracting comprehensive features such as ABCD
features may include some practical limitations like misclassifying melanoma of homogeneous
color and regular shape. Future work on this topic may involve the use of new deep learning
techniques for further improvement of artifacts removal and structures detection. Also, the
hardware design and implementation of the proposed CAD system using high dimensional
Multimedia Tools and Applications

chaotic circuits [44] can be a goal for future investigation. These points are challenges that will
be explored and reported in the near feature.

4 Conclusions

In this paper, a new CAD system is proposed for automated skin lesion classification using
low hardware and software requirements. The proposed approach investigates new pre-
processing techniques for hair removal and lesion identification, Grab-cut in HSV color space
for lesion segmentation, the ABCD rule for discriminating malignant melanoma from benign
lesions, image data augmentation, and pretrained CNN architectures for classification. Differ-
ent pretrained CNN structures such as VGG-16, ResNet50, ResNetX, InceptionV3, and
MobileNet are investigated for skin lesion classification. To investigate the usefulness of the
proposed approach, real skin images covering several cases from the ISIC 2017 and
HAM10000 databases are examined. The use of data augmentation is investigated for
overcoming the problem of data limitation and examining its impact on the performance of
benign-malignant classification for all CNN architectures under investigation. The classifica-
tion performance is identified using several metrics, including AUC, ACC, Se, Pr, F1-score, and
consumed time. Experimental results reveal that the ResNet50 CNN architecture combined
with SVM and data augmentation provides very accurate skin lesion classification with an
average AUC of 99.52%, ACC of 99.87%, Se of 98.87%,Pr of 98.77%, ad F1-score of 7.83%
for the ISIC 2017datasets, and AUC of 97.91%, ACC of 98.79%, Se of 9778%, Pr of 96.86%,
and F1-scor of 97.45% fo the HAM10000 datasets, hile requiring low computational time. The
results highlighted the positive impact of utilizing data augmentation in enhancing the
diagnosis performance. The proposed method is compared with recent skin lesion classifica-
tion systems, and the results elucidate the superior classification results of the proposed method
over other systems under comparison. This demonstrates that the proposed framework can be
used to assist less experienced dermatologists and clinicians in classifying various skin lesions.

Funding Open access funding provided by The Science, Technology & Innovation Funding Authority (STDF)
in cooperation with The Egyptian Knowledge Bank (EKB).

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and
indicate if changes were made. The images or other third party material in this article are included in the article's
Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included
in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or
exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy
of this licence, visit http://creativecommons.org/licenses/by/4.0/.

References

1. Abbas Q, Sadaf M, Akram A (2016) Prediction of dermoscopy patterns for recognition of both melanocytic
and non-melanocytic skin lesions. Computers 5(3):13. https://doi.org/10.3390/computers5030013
2. Albahar MA (2019) Skin lesion classification using convolutional neural network with novel regularizer.
IEEE Access 7:38306–38313. https://doi.org/10.1109/ACCESS.2019.2906241
 Multimedia Tools and Applications

3. Brinker TJ, Hekler A, Utikal JS, Grabe N, Schadendorf D, Klode J, Berking C, Steeb T, Enk AH (2018)
and C. Von Kalle: Skin cancer classification using convolutional neural networks: systematic review. J
Med Internet Res 20(10):e11936. https://doi.org/10.2196/11936
4. Chen B, Lu Y, Pan W, Xiong J, Yang Z, Yan W, Liu L, Qu J (2019) Support vector machine classification
of nonmelanoma skin lesions based on fluorescence lifetime imaging microscopy. Anal Chem 91(16):
10640–10647. https://doi.org/10.1021/acs.analchem.9b01866
5. Dalila F, Zohra A, Reda K, Hocine C (2017) Segmentation and classification of melanoma and benign skin
lesions. Optik 140:749–761. https://doi.org/10.1016/j.ijleo.2017.04.084
6. Demyanov S, Chakravorty R, Abedini M, Halpern A, Garnavi R (2016) Classification of dermoscopy
patterns using deep convolutional neural networks. In: 2016 IEEE 13th International Symposium on
Biomedical Imaging (ISBI). IEEE, Prague, 364–368. https://doi.org/10.1109/ISBI.2016.7493284
7. Eltrass AS, Salama M (2020) Fully automated scheme for computer-aided detection and breast cancer
diagnosis using digitised mammograms. IET Image Process 14(3):495–505. https://doi.org/10.1049/iet-ipr.
2018.5953
8. Feng X, Yao H, Zhang S (2019) An efficient way to refine DenseNet. Signal Image Video Process 13(5):
959–965. https://doi.org/10.1007/s11760-019-01433-4
9. Feng W, Zhang X, Zhao G (2019) ResNetX: a more disordered and deeper network architecture. arXiv
preprint arXiv:1912.12165. Available from: https://arxiv.org/abs/1912.12165. Accessed 20 Jul 2020
10. Garnavi R, Aldeen M, Bailey J (2012) Computer-aided diagnosis of melanoma using border-and wavelet-
based texture analysis. IEEE Trans Inf Technol Biomed 16(6):1239–1252. https://doi.org/10.1109/TITB.
2012.2212282
11. Hardie R, Ali R, Silva D, Kebede TM (2018) Skin lesion segmentation and classification for ISIC 2018
using traditional classifiers with hand-crafted features. arXiv preprint arXiv:1807.07001. Available from:
https://arxiv.org/abs/1807.07001. Accessed 24 Jul 2020
12. Hastie T, Tibshirani R, Friedman J (2009) The elements of statistical learning: data mining, inference and
prediction, 2nd edn. Springer Science & Business Media, Springer-Verlag New York. https://doi.org/10.
1007/978-0-387-84858-7
13. Hosny KM, Kassem MA, Foaud MM (2019) Classification of skin lesions using transfer learning and
augmentation with Alex-net. PLoS One 14(5):e0217293. https://doi.org/10.1371/journal.pone.0217293
14. Howard AG, Zhu M, Chen B, Kalenichenko D, Wang W, Weyand T, Andreetto M, Adam H Mobilenets:
Efficient convolutional neural networks for mobile vision applications. arXiv preprint arXiv:1704.04861,
2017. Available from: https://arxiv.org/abs/1704.04861. Accessed 10 Aug 2020
15. ISIS Archive [electronic resource] (2020) Kitware, Available: Inc. https://isic-archive.com/. Accessed 16
Jan 2020
16. Jaccard P (1912) The distribution of the flora in the alpine zone 1. New Phytol 11(2):37–50. https://doi.org/
10.1111/j.1469-8137.1912.tb05611.x
17. K. Inc (2019) Skin Cancer MNIST: HAM10000. Available: https://www.kaggle.com/kmader/skin-cancer-
mnist-ham10000/version/2. Accessed 24 Jan 2020
18. Kaiming H, Xiangyu Z, Shaoqing R, Jian S (2016) Deep residual learning for image recognition. IEEE
Conf on Computer Vision and Pattern Recognition, Las Vegas, NV, USA. https://doi.org/10.1109/CVPR.
2016.90
19. Kasmi R, Mokrani K (2016) Classification of malignant and benign skin lesions: implementation of
automatic ABCD rule. IET Image Proc 10(6):448–455. https://doi.org/10.1049/iet-ipr.2015.0385
20. Khan MA, Akram T, Sharif M, Shahzad A, Aurangzeb K, Alhussein M, Haider SI, Altamrah A (2018) An
implementation of normal distribution based segmentation and entropy controlled features selection for skin
lesion detection and classification. BMC Cancer 18(1):638. https://doi.org/10.1186/s12885-018-4465-8
21. Khan MA, Sharif M, Akram T, Bukhari SAC, Nayak RS (2020) Developed Newton-Raphson based deep
features selection framework for skin lesion recognition. Pattern Recogn Lett 129:293–303. https://doi.org/
10.1016/j.patrec.2019.11.034
22. Li Y, Zhang J, Gao P, Jiang L, Chen M (2018) Grab cut image segmentation based on image region. In
2018 IEEE 3rd International Conference on Image. Vision and Computing (ICIVC). IEEE, Chongqing,
311–315. https://doi.org/10.1109/ICIVC.2018.8492818
23. Mahbod A, Schaefer G, Wang C, Ecker R, Ellinge I (2019) Skin lesion classification using hybrid deep
neural networks. In: ICASSP 2019-IEEE International Conference on Acoustics, Speech and Signal
Processing (ICASSP). IEEE, Brighton, 1229–1233. https://doi.org/10.1109/ICASSP.2019.8683352
24. Majtner T, Yildirim-Yayilgan S, Hardeberg JY (2016) Combining deep learning and hand-crafted features
for skin lesion classification. In: 2016 Sixth International Conference on Image Processing Theory, Tools
and Applications (IPTA). IEEE, Oulu, 1–6. https://doi.org/10.1109/IPTA.2016.7821017
25. Marcot BG, Hanea AM (2020) What is an optimal value of k in k-fold cross-validation in discrete Bayesian
network analysis? Comput Stat :1–23. https://doi.org/10.1007/s00180-020-00999-9
Multimedia Tools and Applications

26. Matsunaga K, Hamada A, Minagawa A, Koga H (2017) Image classification of melanoma, nevus and
seborrheic keratosis by deep neural network ensemble. arXiv preprint arXiv:1703.03108. https://arxiv.org/
abs/1703.03108. Accessed 12 Aug 2020
27. Menegola A, Fornaciali M, Pires R, Bittencourt FV, Avila S, Valle E (2017) Knowledge transfer for
melanoma screening with deep learning. In 2017 IEEE 14th International Symposium on Biomedical
Imaging (ISBI 2017), pp 297-300, Melbourne, VIC, Australia. https://doi.org/10.1109/ISBI.2017.7950523
28. Monisha M, Suresh A, Bapu BR, Rashmi MR (2019) Classification of malignant melanoma and benign
skin lesion by using back propagation neural network and ABCD rule. Clust Comput 22(5):12897–12907.
https://doi.org/10.1007/s10586-018-1798-7
29. Pan SJ, Yang Q (2009) A survey on transfer learning. IEEE Trans Knowl Data Eng 22(10):1345-1359.
https://doi.org/10.1109/TKDE.2009.191
30. Perez F, Vasconcelos C, Avila S, Valle E (2018) Data augmentation for skin lesion analysis. OR 2.0
context-aware operating theaters, computer assisted robotic endoscopy, clinical image-based procedures,
and skin image analysis. Springer, Cham, pp 303–311. https://doi.org/10.1007/978-3-030-01201-4_33
31. Ramezani M, Karimian A, Moallem P (2014) Automatic detection of malignant melanoma using macro-
scopic images. J Med Signals Sens 4(4):281. https://doi.org/10.4103/2228-7477.144052
32. Rembielak A, Ajithkumar T (2019) Non-melanoma skin cancer–an underestimated global health threat.
Clin Oncol 31(11):735–737. https://doi.org/10.1016/j.clon.2019.08.013
33. Salama MS, Eltrass AS, Elkamchouchi HM (2018) An improved approach for computer-aided diagnosis of
breast cancer in digital mammography. 13th Annual IEEE International Symposium on Medical
Measurements and Applications, Rome, Italy, 1–5. https://doi.org/10.1109/MeMeA.2018.8438650
34. Santos MO (2018) Estimate: cancer incidence in Brazil. Rev Bras Cancerol 64(1):119–120
35. Serra J (1994) Morphological filtering: an overview. Sig Process 38(1):3–11. https://doi.org/10.1016/0165-
1684(94)90052-3
36. Serte S, Demirel H (2019) Gabor wavelet-based deep learning for skin lesion classification. Comput Biol
Med 113:103423. https://doi.org/10.1016/j.compbiomed.2019.103423
37. Shin HC, Roth HR, Gao M, Lu L, Xu Z, Nogues I, Yao J, Mollura D, Summers RM (2016) Deep
convolutional neural networks for computer-aided detection: CNN architectures, dataset characteristics and
transfer learning. IEEE Trans Med Imaging 35(5):1285–1298. https://doi.org/10.1109/TMI.2016.2528162
38. Shorten C, Khoshgoftaar TM (2019) A survey on image data augmentation for deep learning. J Big Data
6(1), 60:1–48. https://doi.org/10.1186/s40537-019-0197-0
39. Simonyan K, Zisserman A (2017) Very deep convolutional networks for large-scale image recognition.
arXiv preprint arXiv:1409.1556. Available from: https://arxiv.org/abs/1409.1556. Accessed 1 Aug 2020
40. Stone M (1977) An asymptotic equivalence of choice of model by cross-validation and Akaike’s criterion. J
R Stat Soc Ser B Methodol 39(1):44–47. https://doi.org/10.1111/j.2517-6161.1977.tb01603.x
41. Szegedy C, Vanhoucke V, Ioffe S, Shlens J, Wojna Z (2016) Rethinking the inception architecture for
computer vision. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,
2818–2826, Las Vegas, NV, USA. https://doi.org/10.1109/CVPR.2016.308
42. Telea A (2004) An image inpainting technique based on the fast marching method. J Graph Tools 9(1):23–
34. https://doi.org/10.1080/10867651.2004.10487596
43. Thanh DN, Prasath VS, Hien NN (2020) Melanoma skin cancer detection method based on adaptive
principal curvature, colour normalisation and feature extraction with the ABCD rule. J Digit Imaging 33:
574–585. https://doi.org/10.1007/s10278-019-00316-x
44. Tsafack N, Kengne J, Abd-El-Atty B, Iliyasu AM, Hirota K, Abd EL-Latif AA (2020) Design and
implementation of a simple dynamical 4-D chaotic circuit with applications in image encryption. Inf Sci
515:191-217. https://doi.org/10.1016/j.ins.2019.10.070
45. Tschandl P, Codella N, Akay NB, Argenziano G, Braun PR, Cabo H, Gutman D, Halpern A, Helba B,
Wellenhof RH, Lallas A (2019) Comparison of the accuracy of human readers versus machine-learning
algorithms for pigmented skin lesion classification: an open, web-based, international, diagnostic study.
Lancet Oncol 20(7):938–947. https://doi.org/10.1016/S1470-2045(19)30333-X
46. Vasconcelos CN, Vasconcelos BN (2017) Experiments using deep learning for dermoscopy image
analysis. Pattern Recogn Lett. https://doi.org/10.1016/j.patrec.2017.11.005
47. Wang Q, Rabab KW (2007) Fast image/video contrast enhancement based on weighted thresholded
histogram equalization. IEEE Trans Consumer Electron 53(2):757–764. https://doi.org/10.1109/TCE.
2007.381756
48. Yoshida T, Celebi ME, Schaefer G, Iyatomi H (2016) Simple and effective pre-processing for automated
melanoma discrimination based on cytological findings. In: IEEE International Conference on Big Data
(Big Data). IEEE, Washington, 3439–3442. https://doi.org/10.1109/BigData.2016.7841005
 Multimedia Tools and Applications

49. Yu L, Chen H, Dou Q, Qin J, Heng PA (2016) Automated melanoma recognition in dermoscopy images via
very deep residual networks. IEEE Trans Med Imaging 36(4):994–1004. https://doi.org/10.1109/TMI.2016.
2642839
50. Yüksel ME, Borlu M (2009) Accurate segmentation of dermoscopic images by image thresholding based on
type-2 fuzzy logic. IEEE Trans Fuzzy Syst 17(4):976–982. https://doi.org/10.1109/TFUZZ.2009.2018300
51. Zachary HR, Secrest AM (2019) Public health implications of google searches for sunscreen, sunburn, skin
cancer, and melanoma in the United States. Am J Health Promot 33(4):611–615. https://doi.org/10.1177/
0890117118811754
52. Zhang X, Zhao S (2018) Segmentation preprocessing and deep learning based classification of skin lesions.
J Med Imaging Health Inform 8(7):1408–1414. https://doi.org/10.1166/jmihi.2018.2448
53. Zhang J, Xie Y, Xia Y, Shen C (2019) Attention residual learning for skin lesion classification. IEEE Trans
Med Imaging 38(9):2092–2103. https://doi.org/10.1109/TMI.2019.2893944
54. Zhang C, Wu X, Gao X (2019) An improved Gaussian mixture modeling algorithm combining foreground
matching and short-term stability measure for motion detection. Multimed Tools Appl, 1–23. https://doi.
org/10.1007/s11042-019-08210-y

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations.