Hindawi
Journal of Food Quality
Volume 2022, Article ID 5845870, 11 pages
https://doi.org/10.1155/2022/5845870
Research Article
An Investigation in Analyzing the Food Quality Well-Being for
Lung Cancer Using Blockchain through CNN
Mohamed Abdelkader Aboamer ,1 Mohamed Yacin Sikkandar ,1 Sachin Gupta ,2
Luis Vives ,3 Kapil Joshi ,4 Batyrkhan Omarov ,5 and Sitesh Kumar Singh 6
1
Department of Medical Equipment Technology, College of Applied Medical Sciences, Majmaah University, Al Majmaah 11952,
Saudi Arabia
2
School of Engineering and Technology, MVN University, Delhi NCR, Haryana, India
3
Peruvian University of Applied Sciences, Lima, Peru
4
UIT, Uttaranchal University, Dehradun, Uttarakhand, India
5
Al-Farabi Kazakh National University, Almaty, Kazakhstan
6
Department of Civil Engineering, Wollega University, Nekemte, Oromia, Ethiopia
Correspondence should be addressed to Sitesh Kumar Singh; sitesh@wollegauniversity.edu.et
Received 21 March 2022; Accepted 13 April 2022; Published 6 May 2022
Academic Editor: Rijwan Khan
Copyright © 2022 Mohamed Abdelkader Aboamer et al. This is an open access article distributed under the Creative Commons
Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properly cited.
Deep learning (DL) is a new approach that provides exceptional speed in healthcare activities with greater accuracy. In this regard,
“convolutional neural network” or CNN and blockchain are two important parts that together fasten the disease detection
procedures securely. CNN can detect and predict diseases like lung cancer and help determine food quality, and blockchain is
responsible for data. This research is going to analyze the extension of blockchain with the help of CNN for lung cancer prediction
and making food safer. CNN algorithm has been trained with a huge number of images by altering the filters, features, epoch
values, padding value, kernel size, and resolution. Subsequently, the CNN accuracy has been measured to understand how these
factors affect the accuracy. A linear regression analysis has been carried out in IBM SPSS where the independent variables selected
are image dataset augmentation, epochs, features, pixel size (90 × 90 to 512 × 512), kernel size (0–7), filters (10–40), and padding.
The dependent variable is the accuracy of CNN. Findings suggested that a larger number of epochs improve the CNN accuracy;
however, when more than 12 epochs are considered, the accuracy may decrease. A greater pixel/resolution also improves the
accuracy of cancer and food image detection. When images are provided with excellent features and filters, the CNN accuracy
improves. The main objective of this research is to comprehend how the independent variables affect the accuracy (dependent),
but the reading may not be fully exact, and thus, the researcher has conceded out a minor task, which delivered evidence
supportive of the analysis and against the analysis. As a result, it can be determined that image augmentation and a large number
of images develop the CNN accuracy in lung cancer prediction and food safety determination when features and filters are applied
correctly. A total of 10–12 epochs are desirable for CNN to receive 99% accuracy with 1 padding.
1. Introduction
The modern-day healthcare industry across the globe has
witnessed the beneficial utilization of machine learning,
artificial intelligence, and blockchain materials for detecting
critical diseases. Previously, detecting proper reasons and
symptoms of critical diseases and predicting fruitful treatment procedures were not easy for healthcare practitioners.
Along with this, the rate of human errors remained a
constant challenge while offering patient care facilities.
Today, the problems can be effectively resolved with the
implementation of blockchain materials in the healthcare
sector. Researchers have identified various effective contributions of blockchain materials through the CNN method
that can be highly used in detecting serious diseases, such as
lung cancer. Convolutional neural networks collect the
2
datasets of the patients suffering from fatal diseases and from
various food products in order to cater the samples using the AI
techniques incorporated with the blockchains that can manipulate the data to discover the serious ailments like lung
cancer and food quality. Scientists and researchers from the
UK, Canada, and Australia have concluded that after conducting relevant surveys. The survey result reflects that around
87% of medical practitioners strongly support the use of
blockchain materials in numerous aspects of patient care [1].
Lung cancer is a slowly growing and serious issue for
patients worldwide, especially for those who are active and
chain smokers. Various healthcare sectors around the globe
now come to a realization that smoking should be banned and
limited in specific areas in order to mitigate the growing
challenges regarding lung cancer. Researchers have identified
that small cells related to lung cancer slowly grow within the
host’s body but spread faster than any other respiratory
disease. Around 70% of people in this fast-growing world are
becoming the prey of lung cancers that need to be properly
diagnosed in order to avoid an excessive death rate across the
globe [2]. From various healthcare surveys and medical reports, researchers are of the opinion that around 55–61% of
people are facing death threats due to lung cancers day by day.
Lung cancer can be determined as the second most common
type of cancer around the world [3]. However, 63% of
physicians have stated that they have to deal with around
600,000 new cases of lung cancer per year [4]. In order to
mitigate the high growth of lung cancer in developed
countries, physicians tend to focus on applying blockchain
materials through the CNN method for offering a better
patient care service. Today, lung cancer has been regarded as
one of the serious death causes of cancer worldwide. Scientists
have determined that in order to deal with around 1.9 million
new lung cancer cases globally, it is important to undertake
the help of blockchain materials for applying the CNN
method [5]. It has been recorded that around 13% of all
critical lung cancer cases can be detected and diagnosed by
using this blockchain technical approach [6]. Lung cancer is a
worldwide burning issue that is increasing rapidly. For that
reason, physicians felt the need for implementing the IoT
facilities for early-stage disease detection of lung cancer. In the
recent years, industries have been pushing toward machine
orientation to make the processes simpler and more efficient
by increasing quality in reduced time. Food quality and safety
are one of the many steps in food processing that is majorly
labor intensive. Artificial intelligence (AI) and deep learning
for determining food classification, quality, and nutrition
have shown their implications in the literature. Their application would increase food industry automation, increase
food safety, and generate higher income through tourism
[7, 8]. Thus, researchers in this research study are going to
investigate the key role played by blockchain materials in
detecting lung cancer and monitoring food quality easier than
before by applying the CNN method.
1.1. Organization. This study has been arranged in a way
that section 1 has discussed the Introduction. Section 2 is
about Literature Survey followed by section 3 that explains
Journal of Food Quality
the Research Methodology. Section 4 elucidates Analysis and
Interpretation followed by section 5 that explains Discussions and Findings, and the section 6 is about Conclusions.
2. Literature Review
Recently, medical science has experienced a huge demand
for applying blockchain materials in various sectors of
healthcare in order to ease disease detection and cure highly
growing issues nowadays. Researchers have analyzed over 1
million cases of lung cancers each year that are highly increasing the death growth rate across the globe due to lung
cancers. The findings from those medical records and cases
at once show that around 3–5% death rate is increasing per
year due to the attack of lung cancer [9]. In order to
comprehend the significance of using blockchain materials,
researchers tend to focus on relevant decisions and opinions
of global physicians for suitable detection of lung cancer in
patients at the initial stage. However, early detection of the
symptoms due to lung cancer is crucial in order to avoid the
high mortality rate to some extent. Doctors and healthcare
practitioners strongly support the utilization of the internet
of things in order to diagnose lung cancer at an early stage.
Researchers have come to a point that the global healthcare
sectors must arrange proper training facilities regarding IoT
and blockchain device implementation Figure 1.
However, the symptoms of lung cancer avoid appearing
until the complete spreading of the disease, which increases
the causes of death among patients. For that reason, physicians undertake effective decisions to detect lung cancer
and predict its symptoms at the early stage by using the
blockchain CNN method in Figures 1 and 2. Physicians have
strongly supported the use of convolution neural networks
as an effective algorithm of deep learning for taking important medical input images [11]. However, with the assistance of blockchain materials, physicians can assign
various learnable biases and weights related to various aspects of lung cancer in those medical images. Thus, by
differentiating among each image, doctors can easily undertake and predict related essential treatment procedures
while curing lung cancer globally. On the contrary, the
artificial CNN applications can be also used widely in
detecting diseases, predicting computer visions, and relevant
clinical image recognition [12]. Proper and accurate image
recognition without any clinical errors can be highly effective for global physicians while detecting harmful factors
related to lung cancer in patients.
Apart from various beneficial impacts on lung cancer
detection, the facilities of blockchain materials also can be
highly used in numerous aspects of healthcare. Among
various growing medical cases, detection of lung cancer and
predicting proper treatment procedures have enhanced the
overall efficiency of the healthcare industry by 57.19% toward sustainable future growth [13]. It has been observed
that blockchain materials can use all the powerful networks
available in different healthcare sectors for exchanging and
preserving previous records and the history of patients’ big
data. Researchers have focused much on the growing applications of blockchain materials in healthcare so that
Journal of Food Quality
3
loT Equipment
CT Scan
Deep neural Network
Traing of neural networks for
detection and classification
Diagnostic desicion
Provided rediologist
Stage 3
Stage 4
Data Collection
Stage 2
Body sensor
Stage 1
Figure 1: Detection of lung cancer by blockchain facilities [10].
Patterns of Local
Contrast
Face
Features
Face
Output Layer
Hidden Layer 2
Hidden Layer 1
Input Layer
Figure 2: Lung cancer prediction by CNN methods [12].
critical errors and fatal diseases can be identified accurately
across the clinical field. On the other hand, blockchain facilities play a major part in dealing with clinical deception in
trials for providing better patient care outcomes.
Researchers while conducting the investigation related to
the use of blockchain in lung cancer detection have identified
a vast range of utilization in relevant healthcare practices. The
CNN method is associated with lung cancer detection while
the technology of the ledger operates all the patients’ medical
history securely across healthcare-based neural networks.
After detecting symptoms and reasons for lung cancer in a
patient’s body, physicians undertake various predictive
measures for better treatment procedures as shown in Figure 3. As a result, they used to apply numerous benefits of
blockchain materials for managing the proper supply of
relevant medicines for lung cancer. However, the practices
Journal of Food Quality
Accuracy
4
Lung Cancer Accuracy vs Data Size
450000
400000
350000
300000
250000
200000
150000
100000
50000
0
1
2
3
Data Size
4
5
Loss
Accuracy
Val Loss
Figure 3: Graph showing lung cancer accuracy prediction by using
blockchain materials [14].
and applications of the blockchain CNN method can at once
aid healthcare researchers to unlock various codes related to
genetics in the sustainable future. While identifying critical
diseases, blockchain also processes overall infrastructure and
network securities. Besides predicting disease symptoms, the
CNN method also can identify patients’ verification and
participants’ authentication regarding studying previous cases
related to lung cancer [14]. Moreover, suitable access to
electronic health information and treatment procedures after
detecting lung cancer can also be conducted with the help of
blockchain materials worldwide.
With the growing concerns about mitigating lung
cancers around the globe, physicians have focused their
medical attention to look out for the key signs of the
disease in a patient. After using blockchain CNN methods,
physicians and scientists can properly classify lung cancer
CT images as shown in Figures 2 and 3. Subsequently,
blockchain materials help doctors to organize all the
knobs of the affected lung to assess all the levels of life risk
and threats. Ilinca thinks, by evaluating numerous parameters of CNN, that the accuracy of time and intricacy
of precision can be highly determined for sensitive lung
tumor cells detection [15]. Blockchain has found its utility
in top global businesses around the world, widely known
to disrupt the world market. Its application in the food
sector has proven advantageous in the following areas
[16, 17]:
(i) Food tampering, fraud, and misrepresentation
(ii) Assisting with large-scale withdrawals of tainted
goods
(iii) Detecting food waste in supply chain operations
(iv) Reducing the likelihood of food spoilage
(v) Allowing businesses to verify sustainably grown or
valid goods source
(vi) Enhancing food tracing, communication, and
collaboration
(vii) Optimizing the process
(viii) Simplifying food marketing operations
The CNN is one such deep learning technology, which
has also been applied to several business sectors, including
the food sector, mainly to identify different food types and
determine their quality by processing multiple food images
across any platform [18]. Combining CNN with blockchain
would only end up increasing the overall efficiency of this
entire process. A blockchain methodical array conducted
through convolution neural networks reflects high usefulness for detecting and diagnosing lung cancer via taking
various breath samples of patients. The sensors presented in
the CNN array can conduct effective discrimination toward
the organic compounds that are volatile and can be traced in
the breath samples of patients [19]. The overall composition
of the blockchain pattern can be efficiently determined after
using CNN method while evaluating real-time symptoms of
exhaled breath for detecting any growth of cancer cells
Figure 4. Nowadays, machine learning with blockchain CNN
technology can easily detect and diagnose lung cancer
symptoms by classifying and recognizing clinical images
through computed tomography Figure 1. Lung cancer detection and treatment have become a trendy topic in
healthcare. Physicians have witnessed various procedures of
automated detection of lung cancer that at once offer numerous advantages in medical fields.
Both the researchers and medical professionals are
showing genuine interest in adapting beneficial approaches
for lung cancer detection and prediction by using blockchain
CNN facilities. On the other hand, the particular CNN
model can enable automatic identification of lung cancer
easier than in past years, which was a highly challenging
practice for doctors. Researchers in the research paper effectively collected and analyzed relevant medical data and
details related to lung cancer by automated blockchain facilities. Besides, the classification of related clinical images
has also become easier in the recent period through computed tomography analysis Figure 4. Nevertheless, there may
be few backlogs in AI while dealing with the healthcare like it
requires human scrutiny, may oversee societal variables, and
may lead to job loss. As per Le and Hsu, the application of
the model with a 98.04% sensitivity level can be highly effective in reducing human errors by 99.56% [20]. Moreover,
with an accuracy rate of 97.23%, researchers can effectively
investigate the use of blockchain facilities in lung cancer
prediction for a sustainable healthcare service in the future.
3. Research Methodology
Physicians and scientists were investigating over the past few
years to detect the proper use of blockchain materials and
IoT in lung cancer detection in order to apply proper
treatments, along with their implementation in the food
sector. Researchers have focused on performing a regression
analysis in this research study for evaluating the contributions of blockchain CNN methods. In this study, researchers
have evaluated the extension of blockchain with the aid of
CNN for lung cancer extrapolation and making food safer.
CNN algorithm has been skilled with a massive number of
images by alterable the filters, features, epoch values, padding value, kernel size, and resolution. Subsequently, the
Journal of Food Quality
X-ray Image Acquisition
5
Image
Preprocessing
Image Enhancement
Lung Regions Extraction
Lung Segmentation
Feature
Extraction
Edge Detection
Analysis of Extracted Lung Region Using Neural Network
Figure 4: Framework for lung cancer detection by applying the
blockchain CNN method [19].
CNN correctness has been measured to comprehend how
these factors disturb the accuracy. A linear regression examination has been carried out in IBM SPSS where the
independent variables selected are image dataset amplifications. While applying a model of regression analysis,
researchers have concentrated more on evaluating numerous linear regression practices for taking suitable decisions
in the sustainable future. On the other hand, various pieces
of medical evidence also have been gathered for recognizing
CT scan images related to lung cancer and noncancer cases
[21]. For exploring the accuracy rate in the CNN outcomes,
researchers tend to minutely analyze all the factors for
regulating the performance of regression analysis toward
easy detection in predicting of lung cancer diseases and
analyzing food quality. Besides conducting a linear regression analysis, researchers also tend to focus on evaluating all
the outcomes of the calculations of descriptive statistics with
both the minimum and maximum values. The entire regression analysis has been outlined with a 95% level of
confidence in the software IBM SPSS. Researchers also have
undertaken various effective decisions while collecting almost 1,00000–2,00000 medical images from healthcare
sectors and food images from social media and restaurants in
order to understand the use of blockchain in lung cancer
detection and food quality. On the other hand, the CNN
accuracy level has also been tested by the researchers against
various augmentations of the clinical image dataset. Apart
from this, numerous lung cancer-related features and epochs
also have been tested and evaluated with the help of the
regression analysis model. Changes and modifications in the
size of images’ pixels (90 × 90 to 512 × 512), and the kernel
size (0–7), have been considered positively for conducting
the overall research method. Researchers have focused much
on analyzing both the findings of primary and secondary
research filters (10–40) and padding. Therefore, to check the
validity of the collected primary results, they have also
conducted secondary research in order to measure the CNN
accuracy as a dependent variable and the rest as independent
variables.
This research has effectively evaluated the significance of
two important variables related to the CNN method toward
focusing more on particular characteristics related to the
early detection of lung cancer. CNN can automatically
recognize and organize features using artificial neural networks with the aid of a trained dataset catered from several
samples of patients with critical issues with almost no intrusion of human efforts. It is also important to examine
various patient-related data including the value of heart rate,
CT scan images, X-ray, and blood pressure data collected
through ML algorithms of IoT [22]. The p value significance
has been successfully evaluated, where the below p value
indicates 0.05 (p < 0.05). However, the value reflects statistical importance related to the particular topic. The entire
analysis also aids researchers in comprehending the importance of the value of Pearson’s correlation, which is
almost close to +/−1. The negative value reflects the accurate
relationship in the research study that has been traced to be
negatively correlated with the variables. Moreover, researchers also explored the storing and sharing of various
medical nodes by using blockchain facilities. The flowchart
for the research followed is shown in Figure 5.
4. Analysis and Interpretation
The regression analysis has been carried out with a 95%
confidence level in IBM SPSS. The accuracy of CNN has been
tested against image dataset augmentation, epochs, features,
pixel size (90 × 90 to 512 × 512), kernel size (0–7), filters
(10–40), and padding. Therefore, the CNN accuracy is a
dependent variable and the rest are independent variables.
Below Table 1 shows the value of the descriptive statistics.
Table 1 shows the minimum value of image data augmentation taken for the experiment, which is 10,000, or the
lowest image set contains 10,000 CT, X-ray scan images of
lungs, and food images. A maximum of 2,00,000 CT, X-ray,
and food images have been taken in the CNN dataset with
the help of blockchain. When the entire testing and training
were over, 99.9% maximum accuracy was observed, and a
minimum of 53.4% accuracy has been obtained. From this
Table 1, it cannot be concluded which factor is responsible
for improving the accuracy of CNN. A maximum of 13 and a
minimum of 1 epoch have been considered. A maximum of
1080 × 1080 and a minimum of 90 × 90 resolution images
have been provided. Among them, mostly 512 × 512 resolution images have been used. A total of 9 kernels have been
considered with 10–40 filters and 0–1 padding.
Table 2 shows the coefficient values, which also define
how the different parameters impact the accuracy of CNN.
In this scenario, blockchain has been used for storing and
collecting data, and then, data were directly transferred to
CNN with blockchain encryption. The t value here shows
how different the parameters impact the accuracy of CNN.
“Significance value” shows whether the variables are statistically significant to each other or not. The image data
augmentation is not statistically significant with CNN accuracy (p > 0.5). The t value, here, is positive (0.641), which
6
Journal of Food Quality
Use of the Block
Chain CNN technology
In lung cancer
detection
Discussion and Findings of
the mixed method
Classification of
numerous medical
opportunities and
challenges
Conclusion
Interpretation from the
analysis of the regression
model
Important Literature
Review related to the use
Of block chain CNN
method in lung cancer
detection
Future Scopes
Regression analysis by
using IBM SPSS
version 6 Software
Mixed methods of
research with 100000200000 clinical images
Figure 5: Research flowchart.
Table 1: Descriptive statistics’ output showing minimum, maximum, mean, and other values.
Valid
N
Missing
Mean
Std.
deviation
Variance
Minimum
Maximum
Statistics
Features
Pixel (x)
20
20
0
0
26.65
471.70
Image data augmentation
20
0
105000.00
Epochs
20
0
7.95
59160.798
3.517
9.533
3500000000.000
10000
200000
12.366
1
13
90.871
10
50
Kernel size
20
0
4.15
Filters
20
0
29.55
Padding
20
0
0.55
Accuracy of CNN
20
0
82.795
312.884
2.412
8.101
0.510
17.3127
97896.537
90
1080
5.818
1
9
65.629
10
40
0.261
0
1
299.731
53.4
99.9
Table 2: Coefficient values.
Coefficients
Model
1
(Constant)
Image data augmentation
Epochs
Features
Pixel (x)
Kernel size
Filters
Padding
t
Sig.
8.609
0.641
6.705
0.618
−1.242
1.570
−1.788
−0.235
0.000
0.534
0.000
0.548
0.238
0.142
0.099
0.818
95.0% confidence interval for B
Lower bound
Upper bound
40.685
68.258
0.000
0.000
3.492
6.855
−0.275
0.492
−0.026
0.007
−0.732
4.511
−1.604
0.158
−5.908
4.756
Dependent variable: Accuracy of CNN.
suggests that when image augmentation increases, the accuracy of CNN increases. As previously stated, a total of
2,00,000 images have been taken for training; thus, when
more images were provided, accuracy improved. However,
the accuracy is not strongly dependent on the image augmentation and number of images (p > 0.5). The accuracy is
strongly dependent on the number of epochs. A higher t
value (6.705) and p < 0.001 suggest that increasing epochs
significantly increases the accuracy of CNN. The number of
features is not statistically significant with the accuracy
(p > 0.5); however, increasing features can increase the accuracy to some extent (t 0.618). The pixels of the images are
showing a negative correlation (−1.242), which suggests that
when pixel size decreases, the accuracy increases. The
Journal of Food Quality
Table 3: ANOVA regression output.
ANOVA
Model
Sum of squares Df Mean square
F
Sig.
Regression
5424.011
7
774.859
34.327 0.000
1 Residual
270.878
12
22.573
Total
5694.889
19
Dependent variable: accuracy of CNN. Predictors: padding, filters, image
data augmentation, features, pixel (x), epochs, and kernel size.
Table 4: Model summary and R values.
Model summaryb
Model
1
R
R square
Adjusted R square
0.976a
0.952
0.925
Std. error of
the estimate
4.7511
a. Predictors: (constant), padding, filters, image data augmentation, features, pixel (x), epochs, kernel size. b. Dependent variable: accuracy of CNN.
Histogram
Dependent Variable: Accuracy of CNN
Mean = 3.46E-15
Std- Dev. = 0.795
N = 20
6
Frequency
relation is not statistically significant although (p > 0.2). As
previously mentioned, mostly, 512 × 512 images have been
used; therefore, the interpretation and analysis are not
accurate.
Kernel size is positively correlated with the CNN accuracy (t 1.570). The regression here is weakly significant
with the accuracy (p > 0.1), which suggests that when kernel
size increases, the accuracy improves. The number of filters
is also slightly significant with the accuracy (p > 0.9) and the
correlation is antiproportional (t −1.788). It suggests that
when the number of filters decreases, the accuracy increases.
Lastly, the padding of CNN is not statistically significant
with the accuracy (p > 0.8). However, the t value suggests
when padding is “0,” the CNN shows higher accuracy, and
when padding is “1,” the accuracy decreases (t −0.235).
Table 3 shows the ANOVA output where it can be
observed that the entire model is statistically significant
(p < 0.001) with an F value of greater than 3.9 (34.327).
Table 4 shows the entire model summary with an adjusted R square value. The R square value suggests the accuracy of the model, which is 0.925. Hence, it can be
suggested that the model is 92.5% accurate (adjusted) with a
4.75% error. Figure 6 suggests that the model is centered and
not skewed in any direction. Figure 7 suggests that the values
are not scattered and the model is statistically significant.
Table 5 and Figure 8 show how the number of epochs
affects the CNN accuracy (positive correlation till 12
epochs).
7
4
2
5. Discussion and Findings
0
-2
-1
0
1
Regression Standardized Residual
2
Figure 6: Residual statistics’ plot showing a centered bell-shaped curve.
Normal P-P Plot of Regression Standardized Residual
Dependent Variable: Accuracy of CNN
1.0
0.8
Expected Cum Prob
The blockchain here is extended for developing connections
with CNN to provide patient information and food statistics.
Initially, CNN was used with AI, but this study is combining
CNN with blockchain, which will ultimately allow the CNN
to classify the image for lung cancer prediction and determine food quality. Blockchain is only responsible for storing
and sharing data between different nodes [23]. Patient
medical data such as heart rate value; computer tomography
or CT scan data (Figure 9); X-ray scanned images; food
images; blood pressure data; oxygen level (using oximeter);
and other essential data are collected by using the internet of
things or IoT devices. These data are further stored in the
blockchain cloud environment and then shared with professionals. Clinicians analyze these data using deep learning
algorithms to predict lung cancer. Similarly, analysts do the
same to check food status and nutrition level.
The current analysis showed that image augmentation
was not statistically significant with the CNN accuracy in
lung cancer prediction. The possible reason is when the
image dataset and augmentation (Figure 10) were increased,
overfitting occurred, which in turn reduced the accuracy
[25]. However, the t value suggested that when image
augmentation and the number of images increase, the CNN
accuracy increases. A similar study has been carried out by
Moreno-Barea et al., which also showed that data augmentation increases the accuracy of deep learning architectures when the dataset is small (less than 1000) [26].
Concerning this, Bandara and other researchers stated that
0.6
0.4
0.2
0.0
0.0
0.2
0.4
0.6
Observed Cum Prob
0.8
1.0
Figure 7: Scattered plot showing the model is statistically
significant.
8
Journal of Food Quality
Table 5: Epochs and accuracy of CNN.
Epochs
Accuracy of CNN
1
55.2
2
53.4
3
55
4
60
5
67
6
68
7
70
8
85
9
95
10
98
12
14
11
98.5
12
98.2
13
96.5
11
99
Accuracy of CNN against Epochs
120
100
80
60
40
20
0
0
2
4
6
8
10
Figure 8: Accuracy of CNN increases with the number of epochs.
000001
000002
000003
000004
000005
000006
000007
000008
000009
000010
000011
000012
000013
000014
000015
000016
000017
000018
000019
000020
000021
000022
000023
000024
000025
000026
000027
000028
000029
000030
000031
000032
Figure 9: CT scan images for CNN training [24].
data augmentation improves the baseline accuracy of CNN
[27]. It has been observed that a large number of epochs
increase the accuracy of CNN. A total of 13 epochs have been
considered here, and when the epoch number is 10–12, the
CNN accuracy was 98–99.9%. A study by Barman et al.
showed that when 11–12 epochs are taken, the CNN shows
maximum accuracy [28]. When 11–12 epochs are taken the
“data passing” is enough; however, when the epochs are
increased, the data become “overfitting” for the CNN architecture. In this case, the researchers observed that 10–12
epochs are enough for obtaining 99% accuracy in lung
cancer detection. The number of features showed no statistically significant relationship with accuracy; however, the
t value was positive. Therefore, the features improved
accuracy by providing the CNN with more accurate details
of CT and X-ray scanned images. A study by Bochkovskiy
et al. showed that a large number of features increase the
accuracy of CNN [29]. Therefore, the analysis and output in
this currency research are reliable.
The pixel size did not show any statistically significant
relationship with the accuracy; however, the t value is
negative. This result seems to be “false negative”; it suggests
that when the resolution of images decreases, the accuracy
increases. However, when resolution increases, the CNN
receives larger pixels and features (details), which are expected to improve the accuracy. The previous paragraph
described that a larger number of features improve CNN
accuracy. To support this sentence, evidence from Borji can
Journal of Food Quality
9
(a)
(b)
(c)
(d)
(e)
Figure 10: Data augmentation techniques [29].
also be observed here, which states that increasing resolution
increases the accuracy. Therefore, the findings from this
current research related to resolution are false and larger
resolution increases accuracy [30].
The kernel size has shown a positive correlation with
lung cancer prediction accuracy. A total of 9 kernels have
been used, and increasing kernel size improved the accuracy
of CNN. Kong and Jang showed that increasing the kernel
increases the accuracy of neural networks [31]. Although
strong evidence has not been obtained regarding this,
however, this study and other available studies showed
kernel size increased the accuracy [32]. In this study, filters
are negatively correlated with accuracy, which suggests that
when more filters are used, it can cause overfitting. Lastly,
0–1 padding has been used and it showed no significant
improvement in the CNN accuracy. However, a study by
Wei and Lin et al. showed 1 padding improved accuracy than
0 padding. In this study, authors have predicted heart disease
using deep learning that still researchers are examining deep
neural networks and CNN [33]. A lot of research has been
conducted in this field, and various researchers have proposed several methodologies, which can be taken as a reference to conduct further research by joining the various
gaps, which lags behind, and further it can be determined
that convolutional neural network or CNN and blockchain
are two important parts that together securely fasten the
disease detection procedures like lung cancer and help in
determining the food superiority with the assistance of
blockchain for numerous data. In this research, we have
strained to analyze the extension of blockchain with the help
of CNN for lung cancer prediction and making food
harmless.
6. Conclusions
This study has been carried out with a larger number of the
training dataset to understand how the number of epochs,
number of images, pixels, features, and padding affect the
CNN accuracy in lung cancer prediction and analyze food
safety. The study found that when the number of epochs is
10–12, the CNN obtained more than 99% accuracy, and
when the epoch number exceeds 12, the accuracy decreased.
A large number of image augmentations improve accuracy
when filters and features are applied correctly; otherwise,
overfitting will decrease the accuracy. The padding did not
show any significant improvement in accuracy; however,
10
after receiving evidence from available studies, it has been
observed that “1” padding improves accuracy than “0”
padding. The entire model is 92.5% accurate with a standard
error of 4.75% because some false results have been obtained.
The primary research has been performed for understanding how the independent variables affect the accuracy
(dependent). The study may not be fully correct, and thus,
the researcher has carried out secondary research, which
provided evidence supportive to the analysis and against the
analysis. Therefore, it can be concluded that image augmentation and a large number of images improve the CNN
accuracy in lung cancer prediction and food safety determination when features and filters are applied correctly. A
total of 10–12 epochs are desirable for CNN to receive 99%
accuracy with 1 padding.
6.1. Future Scopes. The future scopes of CNN and blockchain include efficient image classification and data encryption, respectively. Blockchain has been shown to
improve the value of an organization, and the reports suggest
that the blockchain will help a business to grow faster.
Studies also show that transactions using blockchain are
more secure and safer than banking transactions. Moreover,
due to its minimal compliance cost, the blockchain can
dominate in the future.
On the other hand, CNN, which is a deep learning algorithm, can automatically classify and identify features
without any human intervention. For example, when CT
scan images or X-ray images are provided to the CNN, the
CNN can automatically detect the new features, whereas
other neural networks require human supervision to capture
the features. Moreover, CNN is more effective and efficient
than other architectures. Thus, CNN along with blockchain
technology can provide long-term success to healthcare and
other organizations. However, security is a major concern
that needs to be improved for encrypted operations.
Data Availability
The data shall be made available on request.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
References
[1] A. Jain, A. K. Yadav, and Y. Shrivastava, “Modelling and
optimization of different quality characteristics in electric
discharge drilling of titanium alloy Sheet,” Material Today
Proceedings, vol. 21, pp. 1680–1684, 2019.
[2] P. Ratta, A. Kaur, S. Sharma, M. Shabaz, G. Dhiman, and
A. K. Pandey, “Multiple quality optimizations in electrical
discharge drilling of mild steel sheet,” Material Today Proceedings, vol. 2021, pp. 7252–7261, 2019.
[3] P. Ratta, A. Kaur, S. Sharma, M. Shabaz, and G. Dhiman,
“Application of blockchain and internet of things in
healthcare and medical sector: applications, challenges, and
future perspectives,” in Journal of Food Quality, R. Khan, Ed.,
vol. 2021, Article ID 7608296, 20 pages, 2021.
Journal of Food Quality
[4] D. Ushakov, E. Robu, O. Blagorazumnaia, and S. Kabaha,
“Transnationalized tourism: hyper-advantages from global
competitiveness,” Journal of Environmental Management and
Tourism, vol. 11, no. 6, pp. 1316–1327, 2020.
[5] S. Kumar, W. M. Lim, U. Sivarajah, and J. Kaur, “Artificial
intelligence and blockchain integration in business: trends
from a bibliometric-content analysis,” Information Systems
Frontiers: A Journal of Research and Innovation, vol. 26, no. 2,
pp. 1–26, 2022.
[6] N. Banús, I. Boada, P. Xiberta, P. Toldrà, and N. Bustins,
“Deep learning for the quality control of thermoforming food
packages,” Scientific Reports, vol. 11, no. 1, pp. 21887–21915,
2021.
[7] H. Chen, J. Xu, G. Xiao, Q. Wu, and S. Zhang, “Fast auto-clean
CNN model for online prediction of food materials,” Journal
of Parallel and Distributed Computing, vol. 117, pp. 218–227,
2018.
[8] A. Srinivasulu, K. Ramanjaneyulu, R. Neelaveni et al., “Advanced lung cancer prediction based on blockchain material
using extended CNN,” Applied Nanoscience, pp. 1–13, 2021.
[9] S. Hosseinzadeh Kassani, “Towards secure and intelligent
diagnosis: deep learning and blockchain technology for
computer-aided diagnosis systems,” Doctoral dissertation,
University of Saskatchewan, Saskatoon, Canada, 2021.
[10] P. M. Bala, S. Usharani, G. L. Roselin, A. Balachandar, and
M. Pavithra, “Significance and usage of deep learning in health
care systems: diagnosis of lung cancer,” International Journal
of Recent Development in Computer Technology & Software
Applications, vol. 5, no. 1, 2021.
[11] F. M Bublitz, A. Oetomo, K. S. Sahu et al., “Disruptive
technologies for environment and health research: an overview of artificial intelligence, blockchain, and internet of
things,” International Journal of Environmental Research and
Public Health, vol. 16, no. 20, p. 3847, 2019.
[12] news-medical.net, “Use of blockchain in healthcare,” 2022,
https://www.news-medical.net/health/BlockchainApplications-in-Healthcare.aspx.
[13] N. Ejaz, R. Ramzan, T. Maryam, and S. Saqib, “Big Data
Management of Hospital Data using Deep Learning and
Block-chain Technology: A Systematic Review,” EAI Endorsed
Transactions on Scalable Information Systems, vol. 21, no. 32,
p. e1, 2021.
[14] D. Ilinca, “Applying blockchain and artificial intelligence to
digital health,” in Digital Health Entrepreneurship, pp. 83–101,
Springer, Berlin, German, 2020.
[15] A. Rejeb, J. G. Keogh, S. Zailani, H. Treiblmaier, and K. Rejeb,
“Blockchain technology in the food industry: a review of
potentials, challenges and future research directions,” Logistics, vol. 4, no. 4, p. 27, 2020.
[16] N. Kshetri, “Blockchain and the economics of food safety,” It
Professional, vol. 21, no. 3, pp. 63–66, 2019.
[17] D. J. Attokaren, I. G. Fernandes, A. Sriram, Y. S. Murthy, and
S. G. Koolagudi, “Food classification from images using
convolutional neural networks,” in Proceedings of the TENCON 2017-2017 IEEE Region 10 Conference, pp. 2801–2806,
Penang, Malaysia, November 2017.
[18] S. Paramasivam, C. H. Shen, A. Zourmand, A. K. Ibrahim,
A. M. Alhassan, and A. F. Eltirifl, “Design and modeling of
iotir thermal temperature screening and uv disinfection
sterilization system for commercial application using blockchain technology,” in Proceedings of the 2020 IEEE 10th International Conference on System Engineering and Technology
(ICSET), pp. 250–255, Shah Alam, Malaysia, 2020.
Journal of Food Quality
[19] T. V. L. Tuan-Vinh Le and C. L. H. Tuan-Vinh Le, “A systematic literature review of blockchain technology: security
properties, applications and challenges,” Journal of Internet
Technology, vol. 22, no. 4, pp. 789–801, 2021.
[20] O. L. Saldanha, P. Quirke, N. P. West et al., “Swarm learning
for decentralized artificial intelligence in cancer histopathology,” bioRxiv, vol. 11, 2021.
[21] Y. Xie, L. Lu, F. Gao et al., “Integration of artificial intelligence, blockchain, and wearable technology for chronic
disease management: a new paradigm in smart healthcare,”
Current Medical Science, vol. 41, no. 6, pp. 1123–1133, 2021.
[22] T. Kadir and F. Gleeson, “Lung cancer prediction using
machine learning and advanced imaging techniques,”
Translational Lung Cancer Research, vol. 7, no. 3, pp. 304–312,
2018.
[23] C. J. Lin, S. Y. Jeng, and M. K. Chen, “Using 2D CNN with
Taguchi parametric optimization for lung cancer recognition
from CT images,” Applied Sciences, vol. 10, no. 7, p. 2591,
2020.
[24] C. Shorten and T. M. Khoshgoftaar, “A survey on image data
augmentation for deep learning,” Journal of big data, vol. 6,
no. 1, pp. 60–48, 2019.
[25] A. Borji, “Enhancing sensor resolution improves cnn accuracy
given the same number of parameters or flops,” 2021, https://
arxiv.org/abs/2103.05251.
[26] F. J. Moreno-Barea, J. M. Jerez, and L. Franco, “Improving
classification accuracy using data augmentation on small data
sets,” Expert Systems with Applications, vol. 161, Article ID
113696, 2020.
[27] K. Bandara, H. Hewamalage, Y. H. Liu, Y. Kang, and
C. Bergmeir, “Improving the accuracy of global forecasting
models using time series data augmentation,” Pattern Recognition, vol. 120, Article ID 108148, 2021.
[28] U. Barman, R. D. Choudhury, D. Sahu, and G. G. Barman,
“Comparison of convolution neural networks for smartphone
image based real time classification of citrus leaf disease,”
Computers and Electronics in Agriculture, vol. 177, Article ID
105661, 2020.
[29] A. Gupta and L. K. Awasthi, “Security issues in cross-organizational peer-to-peer applications and some solutions,” in
Communications in Computer and Information Science,
pp. 422–433, Springer, Berlin, German, 2009.
[30] J. B. Kong and M. S. Jang, “Association analysis of convolution
layer, kernel and accuracy in CNN,” The Journal of the Korea
institute of electronic communication sciences, vol. 14, no. 6,
pp. 1153–1160, 2019.
[31] A. Gupta and L. K. Awasthi, “Secure thyself: securing individual peers in collaborative peer-to-peer environments,” in
Proceedings of the 2008 International Conference on Grid
Computing & Applications, pp. 140–146, Las Vegas, NV, USA,
July 2008.
[32] Z. Wei, H. Hu, H. W. Zhou, and A. Lau, “Characterizing rock
facies using machine learning algorithm based on a convolutional neural network and data padding strategy,” Pure
and Applied Geophysics, vol. 176, no. 8, pp. 3593–3605, 2019.
[33] R. Bharti, A. Khamparia, M. Shabaz, G. Dhiman, S. Pande,
and P. Singh, “Prediction of heart disease using a combination
of machine learning and deep learning,” in Computational
Intelligence and Neuroscience, A. A. Abd El-Latif, Ed.,
vol. 2021, Article ID 8387680, 11 pages, 2021.
11