Professional Documents
Culture Documents
A PROJECT REPORT
Submitted By
J.ASHVIK 711219104003
S.D.INIYAN 711219104010
V.SABESAN 711220104703
M.SIVA SURYA 711220104704
Of
BACHELOR OF ENGINEERING
IN
MAY 2023
ANNA UNIVERSITY CHENNAI: 600025
BONAFIDE CERTIFICATE
Place: Avinashipalayam,
Date:
Submitted for the Anna university VIVA-VOCE exam held on
……………………………. at Jai Shriram Engineering college, Tirupur-
638660.
This project would have never been materialized without the innumerable
discussions we had with many people. We would like to extend our sincere
gratitude to all of them for giving meaning and substance to our work
ABSTRACT
TABLE OF CONTENTS
LIST OF FIGURES
LIST OF ABBREVIATIONS
1 INTRODUCTION
1.1overview
1.2. Problems Identified
1.4objective
2 LITERATURE SURVEY
3 SYSTEM ANALYSIS
3.1EXISTING SYSTEM
3.1.1DRAWBACK
3.2PROPOSED SYSTEM
3.2.1ADVANTAGES
4 SYSTEM DESIGN
4.1. System Architecture
4.2. System Flow – Training Phase
5 BLOCK DESCRIPTION
5.2Modules Description
6 SYSTEM REQUIREMENTS
6.1HARDWAR REQUIREMENTS
6.2SOFTWAREREQUIREMENTS
6.3HARDWARE DESCRIPTION
7 RESULTS AND DISCUSSION
8 CONCLUSION& FUTURE
ENHANCEMENT
APPENDIX- A SCREENSHOT
APPENDIX- B SOURCE CODE
REFERENCES
CHAPTER 1
INTRODUCTION
1.1. Overview
The lungs, which is the organ for respiration is a paired cone shaped organs lying in the
thoracic cavity separated from each other by the heart and other structures in the
mediastinum.
Objective:
The aim of this project is a machine learning-based approach using a novel textural feature
descriptor, Shape-dependent Fibonacci-p patterns for effectively distinguishing ASTHMA
DISEASE, viral pneumonia, and normal condition chest radiographs from each other.
Methodology:
This article proposes a machine learning-based approach using a novel textural feature
descriptor, Shape-dependent Fibonacci-p patterns for effectively distinguishing ASTHMA
DISEASE, viral pneumonia, and normal condition chest radiographs from each other. This
descriptor’s key advantage is that it can encrypt textural patterns having different shapes,
orientations, and discontinuities in one operation while inherently removing noise from the
image. Computer simulations for the full radiograph Kaggle dataset show that the proposed
method has better recall performance than the DL methods and the classical Fibonacci
descriptor. Nearly 100% and 98.44% ASTHMA DISEASEdetection accuracy are achieved
for the classification schemes ASTHMA DISEASEvs normal and ASTHMA DISEASEvs
viral pneumonia, respectively. For the lung area-specific radiograph Kaggle dataset, similar
performance was observed for ASTHMA DISEASEdetection for the classification schemes
ASTHMA DISEASEvs normal and ASTHMA DISEASEviral pneumonia. Likewise, for the
COVIDGR dataset, the proposed feature descriptor yielded better performance compared to
most of the DL methods while achieving comparable performance with respect to method
COVID-SDnet. Since the proposed approach is a machine learning model, it does not require
specialized hardware, has less training time, obtains stabilized model with good detection
performance with small training datasets, is lightweight, and can be deployed quickly.
Merits:
Less training time.
Obtains stabilized model with good detection performance with small training
datasets.
Demerits:
Noise error is high.
Poor performance.
2.2. Predictive Modelling of Asthma DiseaseData in the US: Adaptive
Phase-Space Approach
Authors:Vasilis Z. Marmarelis
Year:2020
Link:https://ieeexplore.ieee.org/document/9137691
Objective:
The aim of this project is a novel adaptive methodology for predictive modeling of the time-
course of daily and cumulative confirmed cases of Covid19 has been presented.
Methodology:
There are currently intensified efforts by the scientific community world-wide to analyze the
dynamics of the Asthma Diseasepandemic in order to predict key epidemiological effects and
assist the proper planning for its clinical management, as well as guide sociopolitical
decision-making regarding proper mitigation measures. Most efforts follow variants of the
established SIR methodological framework that divides a population into “Susceptible”,
“Infectious” and “Recovered/Removed” fractions and defines their dynamic inter-
relationships with first-order differential equations. This article proposes a novel approach
based on data-guided detection and concatenation of infection waves – each of them
described by a Riccati equation with adaptively estimated parameters. Methods: This
approach was applied to Asthma Diseasedaily time-series data of US confirmed cases,
resulting in the decomposition of the epidemic time-course into five “Riccati modules”
representing major infection waves to date (June 18th). Results: Four waves have passed the
time-point of peak infection rate, with the fifth expected to peak on July 20th. The obtained
parameter estimates indicate gradual reduction of infectivity rate, although the latest wave is
expected to be the largest.
Merits:
Accuracy is high.
Performance is high.
Demerits:
Efficiency is low.
High cost.
2.3. Unsupervised Anomaly Detection in Multivariate Spatio-Temporal
Data Using Deep Learning: Early Detection of ASTHMA
DISEASEOutbreak in Italy
Authors:Yildiz Karadayi; Mehmet N. Aydin; Arif Selçuk Öǧrencí
Year: 2020
Link:https://ieeexplore.ieee.org/document/9187620
Objective:
The aim of this project is to a hybrid deep learning framework is proposed to solve the
unsupervised anomaly detection problem in multivariate spatio-temporal data for ASTHMA
DISEASEdetection.
Methodology:
Unsupervised anomaly detection for spatio-temporal data has extensive use in a wide variety
of applications such as earth science, traffic monitoring, fraud and disease outbreak detection.
Most real-world time series data have a spatial dimension as an additional context which is
often expressed in terms of coordinates of the region of interest (such as latitude - longitude
information). However, existing techniques are limited to handle spatial and temporal
contextual attributes in an integrated and meaningful way considering both spatial and
temporal dependency between observations. This article, a hybrid deep learning framework is
proposed to solve the unsupervised anomaly detection problem in multivariate spatio-
temporal data. The proposed framework works with unlabelled data and no prior knowledge
about anomalies are assumed. As a case study, use the public ASTHMA DISEASEdata
provided by the Italian Department of Civil Protection. Northern Italy regions’ ASTHMA
DISEASEdata are used to train the framework; and then any abnormal trends or upswings in
ASTHMA DISEASEdata of central and southern Italian regions are detected. The proposed
framework detects early signals of the ASTHMA DISEASEoutbreak in test regions based on
the reconstruction error. For performance comparison, perform a detailed evaluation of 15
algorithms on the COVID19 Italy dataset including the state-of-the-art deep learning
architectures. Experimental results show that our framework shows significant improvement
on unsupervised anomaly detection performance even in data scarce and high contamination
ratio scenarios (where the ratio of anomalies in the data set is more than 5%). It achieves the
earliest detection of ASTHMA DISEASEoutbreak and shows better performance on tracking
the peaks of the ASTHMA DISEASEpandemic in test regions. As the timeliness of detection
is quite important in the fight against any outbreak, our framework provides useful insight to
suppress the resurgence of local novel coronavirus outbreaks as early as possible.
Merits:
It shows better performance on tracking the peaks of the ASTHMA
DISEASEpandemic in test regions.
Achieve early detection.
Significant improvement on unsupervised anomaly detection performance even in
data scarce and high contamination ratio scenarios.
Demerits:
Too expensive.
Underlying system in the presence of noise and anomaly which pose extra
difficulty.
2.4. CovidGAN: Data Augmentation Using Auxiliary Classifier GAN for
Improved Asthma DiseaseDetection
Authors:Abdul Waheed; Muskan Goyal; Deepak Gupta; Ashish Khanna; Fadi Al-
Turjman; Plácido Rogerio Pinheiro
Year:2020
Link:https://ieeexplore.ieee.org/document/9093842
Objective:
The aim of this project is data augmentation method Using Auxiliary Classifier GAN for
Improved Asthma DiseaseDetection.
Methodology:
This analysis still has a variety of limitations. Firstly, GAN architecture and training can be
improved further. Secondly, it used a small dataset because of the time constraints and
difficulty in gathering enough data. The quality of the synthetic samples produced in this
research could be improved by integrating more labelled data which improves the learning
process of GAN. Thirdly, the dataset is obtained from various sources and cross-centre
validations were not conducted in this analysis. This article has made every effort to ensure
that the data collected is correctly labelled. Any mistake in data labelling, however, would
probably affect the results reported. Such an impact could be especially pronounced when the
dataset is small. Lastly, the only way to reliably detect ASTHMA DISEASEis through
medical assistance and clinical testing. The findings of this article provide promising results
that encourage the use of this approach to make more robust radiology systems. This article
also promotes a systematic large-scale gathering of COVID-CXR images.
Merits:
Accuracy is high.
Time complexity is low.
Demerits:
Substantial amount of training data.
High cost and poor performance.
2.5. ASTHMA DISEASEDetection Based on Image Regrouping and
Resnet-SVM Using Chest X-Ray Images
Authors:Changjian Zhou; Jia Song; Sihan Zhou; Zhiyao Zhang; Jinge Xing
Year: 2020
Link:https://ieeexplore.ieee.org/document/9446895
Objective:
The aim of this project is a novel chest X-ray images based ASTHMA DISEASEdetection
method, the image regrouping and the residual encoder block combined for feature learning
and extraction, the extracted features were input into SVM for classification.
Methodology:
This article proposed a novel chest X-ray images based ASTHMA DISEASEdetection
method, the image regrouping and the residual encoder block combined for feature learning
and extraction, the extracted features were input into SVM for classification. The proposed
method made a concerted effort to reduce the interference of background and skeleton noise,
which showed strong feature learning and expression ability by combined between deep
residual network and Support Vector Machine. The proposed method investigated the
potential biomarkers in the CXR images and found the random localized distributed in lung
region and the noise such as background, spine and ribs have interference of classifiers. To
address this problem, divided the lung region into 15 × 15px pieces, and randomly regrouped
them in to a new image; the regrouped image kept the original ASTHMA DISEASEfeatures
and weakened the background and ribs noises. This article also proposed a residual encoder
block for feature learning, which inspired from auto-encoder. The purposed of auto-encoder
is learning a strong feature for image reconstruction; rather, concentrated on more feasible
work such as feature classification to detection ASTHMA DISEASE. Since the deep residual
network had achieved the state-of-art in classification but need large-scale training data. As
this reason, extend the residual network for feature extraction and input the extracted features
into SVM for classification.
Merits:
More feasible.
Strong feature learning and expression ability.
Demerits:
Lower accuracy.
Efficiency is low.
2.6. Deep learning for mycoplasma pneumonia discrimination from
pneumonias like ASTHMA DISEASE
Authors:Ali Serener; Sertan Serte
Year: 2020
Link:https://ieeexplore.ieee.org/document/9254561
Objective:
The aim of this project is to lessen these diagnosis difficulties by using several deep learning
methods on computed tomography (CT) images to classify them as having mycoplasma
pneumonia, typical viral pneumonia or ASTHMA DISEASE.
Methodology:
Mycoplasma pneumonia, also known as atypical pneumonia, is a contagious respiratory
infection caused by the bacterium mycoplasma pneumonia. It spreads through respiratory
droplets that contain the bacteria which can cause tracheobronchitis, upper respiratory tract
disease as well as pneumonia. Common symptoms are dry cough, fever, shortness of breath,
and fatigue. Viral pneumonia, on the other hand, is a lung infection caused by viruses such as
influenza virus, rhinovirus, and coronavirus. It is acquired by breathing or touching and
results in lung air sucks being infected and filled up with fluid. Common signs of viral
pneumonia are cough, fever, shortness of breath, and loss of appetite. If the lungs are also
invaded with bacteria, bacterial pneumonia symptoms might also occur.This article uses deep
learning techniques on CT images to help differentiate mycoplasma pneumonia from other
types of pneumonias. Specifically, it ran experiments involving deep learning on CT images
to distinguish between mycoplasma pneumonia and ASTHMA DISEASE. It also
experimented and analysed the discernment of mycoplasma pneumonia from typical viral
pneumonia. Using seven different deep learning architectures, it analysed the detection
performance of these discernments. The results show ResNet-18 and MobileNet-v2
architectures as good choices to employ in order to distinguish mycoplasma pneumonia from
typical viral pneumonia and ASTHMA DISEASE.
Merits:
It greatly distinguishes mycoplasma pneumonia from typical viral pneumonia and
ASTHMA DISEASE.
Accuracy is high.
Demerits:
Poor performance.
Low accuracy.
2.7. Texture Analysis of Ultrasound Images for Pneumonia Detection in
Pediatric Patients
Authors:S. L. Contreras-Ojeda; C. Sierra-Pardo; J. A. Dominguez-Jimenez; J. Lopez-
Bueno; S. H. Contreras-Ortiz
Year: 2019
Link:https://ieeexplore.ieee.org/document/8730238
Objective:
The aim of this project is an approach for pneumonia detection based on texture analysis of
ultrasound images.
Methodology:
Pneumonia is a condition that can be life-threatening and affects a high number of children
around the world. Lung ultrasound can be used for the diagnosis of pneumonia, but requires
high experience. This work suggests that pneumonia detection is possible using texture
analysis of ultrasound images. Differences at the mean levels were found in the mean and
median. The pixels of images with pneumonia had lower intensity than the healthy ones. The
results suggest that texture image processing of lung ultrasound can aid in the diagnosis of
pneumonia in pediatric patients. However, it should be noted that pixel intensity is sensible to
the scanner gain settings, firm placement of the transducer, and artifacts such as acoustic
shadowing. Therefore, it is important to look for other metrics to better characterize tissues
with pneumonia.
Merits:
High accuracy.
Speed is high to get faster result.
Demerits:
Requires high experience.
Not good specificity for automatic pneumonia detection.
2.8. Field Trial of Aspiration Pneumonia Predicion based on Electronic
Medical Records
Authors:Masahiro Hayashitani; Eiji Yumoto; Toshinori Hosoi; Masahiro Kubo; Hiroyuki
Hiramitsu; Kensuke Kato; Mayumi Moriguchi
Year: 2019
Link:https://ieeexplore.ieee.org/document/8904811
Objective:
The aim of this project is prediction method of aspiration pneumonia in department of
neurosurgery in order to reduce workload about care for aspiration pneumonia.
Methodology:
In Japan, the lack of labour of medical staff causes long working hours. In particular, the
increase in care time due to the onset of complications is an issue in Kitahara International
Hospital (KIH), which mainly has neurosurgery. Among the complications, aspiration
pneumonia occurs in elderly stroke patients. In our estimation, medical staff have to provide
post-onset care four hours a day per patient of aspiration pneumonia. Preventive cares like
oral care and breathing training are provided in order to avoid the onset of aspiration
pneumonia. However, it takes long time to provide the preventive cares for many patients.
This article proposed the prediction method of aspiration pneumonia in department of
neurosurgery in order to reduce workload about care for aspiration pneumonia. The
prediction method is based on electronic medical records including age, sex, and vital signs.
This article demonstrated the method in KIH. In the field trial, the medical staff provided
preventive cares based on the output of method. It showed that the trial reduced care time of
the medical staff by 10%, and the number of patients with aspiration pneumonia was reduced.
Merits:
Number of patients with aspiration pneumonia was reduced.
Reduced care time of the medical staff by 10%.
Demerits:
Heavy workload about care for aspiration pneumonia.
Low accuracy.
2.9. PNet: An Efficient Network for Pneumonia Detection
Authors:Zhongliang Li; Juan Yu; Xuechen Li; Yingqi Li; Weicai Dai; Linlin Shen; Lisha
Mou; Zuhui Pu
Year:2019
Link:https://ieeexplore.ieee.org/document/8965660
Objective:
The aim of this project is to a deep learning-based framework, PNet, for pneumonia
detection.
Methodology:
Pneumonia is a common lung disease and affects millions of people worldwide each year.
The chest X-ray is one of the most commonly accessible radiological examinations for
screening and diagnosing lung diseases. Usually, chest X-ray images (CXRs) are taken at
physical examination and evaluated by radiologists. The unbalance between tremendous
numbers of CXRs and limited number of radiologists need to be solved by Computer Assisted
Diagnosis (CAD). As deep neural networks have shown promising results in CAD,
implemented a deep learning-based framework, PNet, for pneumonia detection. 10784 chest
X-ray images collected at the Shenzhen No.2 People’s Hospital were employed for training
and evaluation. The experimental results showed that the proposed PNet (1,695,777
parameters) achieved higher accuracy and F1 score than classical networks like AlexNet
(42,725,889 parameters) and VGG16 (27,560,769 parameters).
Merits:
Higher accuracy.
By using small size convolution filters to extract image features is good.
Demerits:
Unbalance between tremendous numbers of CXRs.
Limited number of radiologists.
2.10. Deep Online Sequential Extreme Learning Machines and its
Application in Pneumonia Detection
Authors:Sriram Vijendran; Rahul Dubey
Year:2019
Link:https://ieeexplore.ieee.org/document/8710700
Objective:
The aim of this project is an ELM based comprehensive adaptive image classifier for the
detection of pneumonia by the use of Chest X-Rays for faster diagnosis and earlier delivery
of treatment.
Methodology:
Deep neural networks have demonstrated high levels of accuracy in the fields of image
classification. Deep learning is a multilayer perceptron artificial neural network algorithm,
that uses a backpropagation based learning technique to approximate complicated functions
and alleviating the difficulty associated with optimizing deep models. Multilayer extreme
learning machine (MLELM) is a learning algorithm of an artificial neural network which
takes advantages of deep learning and extreme learning machine. Not only does MLELM
approximate the complicated function but it also does not need to iterate during the training
process. Furthermore, Online Sequential Extreme Learning Machines (OSELM) is an
adaptive algorithm based on ELM that does not require fresh training when faced with a new
dataset, but can adapt to the new dataset by being trained on the new dataset alone.This
article presents the potential application of Deep OS-ELM’s in the field of medical imaging
studies, specifically in the detection of pneumonia from Chest XRays. Detailed modelling
and analysis of images are presented. The proposed model has shown better performance than
SVM’s and ELM’s in terms of accuracy and better training time than CNN’s. Test results
indicate that the proposed model is highly effective in the diagnosis of pneumonia through
Chest X-Rays and can be implemented easily with minimal computational requirements.
Merits:
Its effectiveness is high.
Accuracy is high.
Demerits:
Time consuming process.
It requires proper tuning hyper parameters such as learning rates, number of epochs,
stopping criteria etc.
CHAPTER 3
SYSTEM ANALYSIS
3.1. Existing System
The recently proposed methods for detection/classification of lung diseases such as Asthma
Disease, viral pneumonia, and bacterial pneumonia utilized the machine learning models
primarily based on SVM, K-NN, Logistic Regression, ANN, Random Forest and Ensemble
Learning. However, differentiating the Asthma Diseasecaused pneumonia from viral and
bacterial pneumonia is still a challenging research problem due to lack of the sufficient
experiments on scalable datasets. The following research gaps of recent methods for Asthma
Diseaseand pneumonia disease detection using X-ray scans are the key motivations for the
proposed method.
Machine Learning is an art where the machines are made to learn by some given data and so-
called answers, and not by programming.
K-Nearest Neighbours Classifier
K Nearest Neighbours classifier is Supervised classification method. During the training
time, without having any previous information about the structure of the dataset it classifies
the training set based on the k value. In the proposed work Scikit-learn python library used
for importing the KNN classifier in the program. K Nearest Neighbours algorithm known as a
lazy learner as well because it doesn’t learn from the training data but memorize the training
dataset. The algorithm is based on similar features. As there is no training phase, K Nearest
Neighbours keeps all training data and used it during the testing phase. How closely our
features resemble out training set determines how images classify a given data point. The
outcome of this algorithm is a discrete value which is based on the majority of votes from its
neighbours. If the k value rises the confidence in prediction improves
Artificial Neural Networks (ANN)
An ANN is a computational model based on the human Lung. ANNs contain some nodes
which are connected through weights. Each node receives data from previous nodes, adds it
together and outputs data through a nonlinear function, and then propagates data to
proceeding nodes. The first neuronal layer of the ANN is the input layer composed of
variable number of collected data from observation. The next additional neuronal layers
compose of hidden layers created to generate a variable number of numerical combinations.
The last neuronal layer named output layer generates the answer (numbers that represent the
output). Most of ANN models are based on the idea of supervised training. There are two
phases in the ANN action: training/ learning phase and test phase. A validating phase is still
often made. In training/learning phase, input data are presented to the ANN and weights are
adjusted and fixed. In other words, the ANN does indeed learn the input patterns in the
learning phase.
Support Vector Machines (SVMs)
SVM is a supervised machine learning algorithm that predicts an optimal hyperplane in an n-
dimensional space to divide the training set into multiple classes. Different kernel functions
can be specified for the decision function depending on the problem. They could be used to
implement a multi-class classification on a dataset, providing in advance a subset of labelled
data needed for model training.
Random Forest
Random Forest Algorithm is widely used because of its capability to solve both classification
and regression functions and its simplicity. It is a kind of supervised learning algorithm in
which many decision trees combine to give a more stable and accurate precision. For large
datasets, this algorithm prevents the overfitting of the data by building small decision trees by
creating a subsample space of random features. Whilst there can be some complex
computations resulting in more training time, the results produced are good because of
combinations of these smaller trees. The major obstacle of random forest is that a huge
variety of trees could make the algorithm too sluggish and useless for real-time predictions.
In general, those algorithms are rapid to train, however pretty gradual to create predictions as
soon as
they’re trained. An extra correct prediction calls for extratrees, which results in a slower
model. In maximum real global applications, the random forest is rapidly sufficient however
there can genuinely be conditions wherein runtime overall performance is critical and
different methods
might be preferred.
Ensemble Methods
o XGBoost
XGBoost is a decision-tree-based ensemble Machine Learning algorithm that uses a gradient
boosting framework. In prediction problems involving unstructured data (images, text, etc.)
artificial neural networks tend to outperform all other algorithms or frameworks.
o Bagging
Bootstrap aggregating also called bagging (from bootstrap aggregating), is a machine
learning ensemble meta-algorithm designed to improve the stability and accuracy of machine
learning algorithms used in statistical classification and regression. It also reduces variance
and helps to avoid overfitting.
o ADABoost
AdaBoost algorithm, short for Adaptive Boosting, is a Boosting technique that is used as an
Ensemble Method in Machine Learning. It is called Adaptive Boosting as the weights are re-
assigned to each instance, with higher weights to incorrectly classified instances.
3.2. Disadvantages
Approximately 95% of ML-based methods were failed to consider the challenges of
X-ray image quality enhancement. Therefore, infected areas of X-ray images were not
accurately identified during automatic feature extraction.
Existing ML-based works took complete lung image for automatic feature extraction,
but only the features of infected lung regions are relevant for diagnostics. The lack of
ROI estimation in chest X-ray images leads to high-dimensional and irrelevant feature
extraction for classification. It also restricts the severity analysis of disease due to the
lack of ROI-specific features.
High training time of ML is a challenging issue that leads to a computationally
inefficient solution for early detection of lung diseases.
Asthma Diseaseand pneumonia detection methods using Machine learning have been
evaluated on small X-ray samples in which 10–15% samples were considered for
testing and 80–85% samples were used for training and validation. To claim
efficiency and reliability, such models require a better training and testing ratio.
The automatic feature extraction approaches of ML relied on pre-training models
using irrelevant ImageNet datasets for lung disease predictions.
As the Asthma Diseasecaused pneumonia is relatively new, lack of sufficient medical
data in ML pre-trained models led to unreliable feature extraction.
3.3. Proposed System
Deep Convolutional Neural Network has been proposed for the early detection, segmentation,
and diagnosis of chronic renal failure on the Internet of Medical Things (IoMT) platform.In
this project, a deep learning-based Deep Convolutional Neural Network (DCNN) method has
been proposedfor the early detection, segmentation, and diagnosis of Covid 19 or Pneumonia
or Normal. In the proposed DCNN system, the noise mustbe reduced after acquiring a raw
image, as it can be an importantsegmentation problem. The pre-treatment technique based on
awavelet is used to minimize noise. The denoted RPN image segmentsthe lung area and
extracts more processing features of characteristicsthat have been chosen to indicate an
abnormality inthe lung disease. A graphical userinterface is developed for manual
segmentation to extract theregion of lung area.
3.3.1. Region Proposal Network
This region proposal network takes convolution feature map that is generated by the
backbone layer as input and outputs the anchors generated by sliding window convolution
applied on the input feature map.
CHAPTER 4
SYSTEM DESIGN
4.1. System Architecture
Training Phase Testing Phase
Classification Localization
Clustering
Preprocessing
ResDep Doctor
ar
Server Admin
The procedure to measure PPM using MQ sensor is the same but few constant values will
vary based on the type of MQ sensor used. Basically, we need to look into the (Rs/Ro) VS
PPM graph given in the datasheet (also shown below).
The value of Ro is the value of resistance in fresh air and the value of Rs is the value of
resistance in Gas concentration. First, you should calibrate the sensor by finding the values of
Ro in fresh air and then use that value to find Rs using the formulae
Once we calculate Rs and Ro we can find the ratio and then use the graph shown above we
can calculate the equivalent value of PPM for that particular gas.
MQ-135
The MQ135 gas sensor has high sensitivity in ammonia, sulphide, benzene steam, smoke, and
in other harm full gas. It is low cost and suitable for different applications.
The above figure shows the typical sensitivity characteristics of the MQ-135 for several
gases. in their: Temp: 20, Humidity: 65%, O2 concentration 21%, RL=20kΩ,
Ro: sensor resistance at 100ppm of NH3 in the clean air.
Rs: sensor resistance at various concentrations of gases.
The value of Ro is the value of resistance in fresh air (or the air with we are comparing) and
the value of Rs is the value of resistance in Gas concentration. First you should calibrate the
sensor by finding the values of Ro in fresh air and then use that value to find Rs using the
below formula:
Once we calculate Rs and Ro we can find the ratio and then using the graph shown above we
can calculate the equivalent value of PPM for that particular gas.
MQ-137
The MQ-137 is an Ammonia (NH₃) gas sensor. The sensing element is SnO2, which has
lower conductivity in clean air. When NH₃ (Ammonia) gas exists, the sensor’s conductivity
gets higher along with the gas concentration rising. A simple circuit makes measuring this
change in conductivity and turning it into data fairly straight-forward, but does require some
calibration.
The procedure to measure PPM using MQ sensor is the same but few constant values will
vary based on the type of MQ sensor used. Basically, we need to look into the (Rs/Ro) VS
PPM graph given in the datasheet (also shown below).
The value of Ro is the value of resistance in fresh air and the value of Rs is the value of
resistance in Gas concentration. First you should calibrate the sensor by finding the values of
Ro in fresh air and then use that value to find Rs using the formula:
Once we calculate Rs and Ro we can find the ratio and then using the graph shown above we
can calculate the equivalent value of PPM for that particular gas.
MQ-138
MQ138 gas sensor has a high sensitivity to Toluene, Acetone, Ethanol, and Formaldehyde,
also to other organic steam. The sensor could be used to detect different organic steam, it is
with low cost and suitable for different application.
The 14-digital input/output pins can be used as input or output pins by using pinMode (),
digitalRead () and digitalWrite () functions in Arduino programming. Each pin operates at 5V
and can provide or receive a maximum of 40mA current, and has an internal pull-up resistor
of 20-50 KOhms which are disconnected by default. Out of these 14 pins, some pins have
specific functions as listed below:
Serial Pins 0 (Rx) and 1 (Tx): Rx and Tx pins are used to receive and transmit TTL
serial data. They are connected with the corresponding ATmega328P USB to TTL
serial chip.
External Interrupt Pins 2 and 3: These pins can be configured to trigger an interrupt
on a low value, a rising or falling edge, or a change in value.
PWM Pins 3, 5, 6, 9 and 11: These pins provide an 8-bit PWM output by using
analogWrite() function.
SPI Pins 10 (SS), 11 (MOSI), 12 (MISO) and 13 (SCK): These pins are used for
SPI communication.
In-built LED Pin 13: This pin is connected with a built-in LED, when pin 13 is
HIGH – LED is on and when pin 13 is LOW, its off.
Along with 14 Digital pins, there are 6 analog input pins, each of which provide 10 bits of
resolution, i.e., 1024 different values. They measure from 0 to 5 volts but this limit can be
increased by using AREF pin with analog Reference () function.
Analog pin 4 (SDA) and pin 5 (SCA) also used for TWI communication using Wire
library.
Arduino Uno has a couple of other pins as explained below:
AREF: Used to provide reference voltage for analog inputs with analogReference ()
function.
Reset Pin: Making this pin LOW, resets the microcontroller.
CHAPTER 6
SYSTEM SPECIFICATION
6.1 Hardware specification
Processors: Intel® Core™ i5 processor 4300M at 2.60 GHz or 2.59 GHz (1
socket, 2 cores, 2 threads per core), 8 GB of DRAM
Disk space: 320 GB
Operating systems: Windows® 10, macOS*, and Linux*
6.2 Software specification
PHP 5 or Python 3.2
MySQL
WAMP Server 2.0
Macromedia Dreamviewer 8 IDE
Embedded C
Arduino IDE
6.3. Smart Bandage Unit
Arduino UNO.
Level Sensor
Nano GPS
Esp8266
OLED Display.
What is Arduino?
Twitter message - and turn it into an output - activating a motor, turning on an LED,
publishing something online. You can tell your board what to do by sending a set of
Over the years Arduino has been the brain of thousands of projects, from everyday objects to
artists, programmers, and professionals - has gathered around this open-source platform, their
Arduino was born at the Ivrea Interaction Design Institute as an easy tool for fast prototyping,
reached a wider community, the Arduino board started changing to adapt to new needs and
challenges, differentiating its offer from simple 8-bit boards to products for IoT applications,
wearable, 3D printing, and embedded environments. All Arduino boards are completely
open-source, empowering users to build them independently and eventually adapt them to
their particular needs. The software, too, is open-source, and it is growing through the
Why Arduino?
Thanks to its simple and accessible user experience, Arduino has been used in thousands of
different projects and applications. The Arduino software is easy-to-use for beginners, yet
flexible enough for advanced users. It runs on Mac, Windows, and Linux. Teachers and
students use it to build low-cost scientific instruments, to prove chemistry and physics
principles, or to get started with programming and robotics. Designers and architects build
interactive prototypes, musicians and artists use it for installations and to experiment with
new musical instruments. Makers, of course, use it to build many of the projects exhibited at
the Maker Faire, for example. Arduino is a key tool to learn new things. Anyone - children,
hobbyists, artists, programmers - can start tinkering just following the step-by-step
instructions of a kit, or sharing ideas online with other members of the Arduino community.
There are many other microcontrollers and microcontroller platforms available for physical
computing. Parallax Basic Stamp, Netmedia's BX-24, Phidgets, MIT's Handy board, and
many others offer similar functionality. All of these tools take the messy details of
simplifies the process of working with microcontrollers, but it offers some advantage for
Writing Sketches
Programs written using Arduino Software (IDE) are called sketches. These sketches are
written in the text editor and are saved with the file extension. Ino. The editor has features for
cutting/pasting and for searching/replacing text. The message area gives feedback while
saving and exporting and also displays errors. The console displays text output by the
Arduino Software (IDE), including complete error messages and other information. The
bottom righthand corner of the window displays the configured board and serial port. The
toolbar buttons allow you to verify and upload programs, create, open, and save sketches, and
pde. It is possible to open these files with version 1.0, you will be prompted to save the
Include Library
Adds a library to your sketch by inserting #include statements at the start of your
code. For more details, see libraries below. Additionally, from this menu item you can
access the Library Manager and import new libraries from .zip files.
Add File...
Adds a source file to the sketch (it will be copied from its current location). The new
file appears in a new tab in the sketch window. Files can be removed from the sketch
using the tab menu accessible clicking on the small triangle icon below the serial
Tools
Auto Format
This formats your code nicely: i.e. indents it so that opening and closing curly braces
line up, and that the statements inside curly braces are indented more.
Archive Sketch
Archives a copy of the current sketch in .zip format. The archive is placed in the same
Fixes possible discrepancies between the editor char map encoding and other
Serial Monitor
Opens the serial monitor window and initiates the exchange of data with any
connected board on the currently selected Port. This usually resets the board, if the
Select the board that you're using. See below for descriptions of the various boards.
Port
This menu contains all the serial devices (real or virtual) on your machine. It should
automatically refresh every time you open the top-level tools menu.
Programmer
For selecting a hardware programmer when programming a board or chip and not
using the onboard USB-serial connection. Normally you won't need this, but if
Here you find easy access to a number of documents that come with the Arduino Software
(IDE). You have access to Getting Started, Reference, this guide to the IDE and other
documents locally, without an internet connection. The documents are a local copy of the
Find in Reference
This is the only interactive function of the Help menu: it directly selects the relevant
page in the local copy of the Reference for the function or command under the cursor.
Sketchbook
The Arduino Software (IDE) uses the concept of a sketchbook: a standard place to store your
programs (or sketches). The sketches in your sketchbook can be opened from the File >
Sketchbook menu or from the Open button on the toolbar. The first time you run the Arduino
software, it will automatically create a directory for your sketchbook. You can view or
Beginning with version 1.0, files are saved with a .ino file extension. Previous versions use
the. pde extension. You may still open. pde named files in version 1.0 and later, the software
Allows you to manage sketches with more than one file (each of which appears in its own
tab). These can be normal Arduino code files (no visible extension), C files (.c extension), C+
Uploading
Before uploading your sketch, you need to select the correct items from the Tools >
Board and Tools > Port menus. The boards are described below. On the Mac, the serial port
or COM4, COM5, COM7, or higher (for a USB board) - to find out, you look for USB serial
device in the port section of the Windows Device Manager. On Linux, it should
be /dev/ttyACMx, /dev/ttyUSBx or similar. Once you've selected the correct serial port and
board, press the upload button in the toolbar or select the Upload item from the Sketch menu.
Current Arduino boards will reset automatically and begin the upload. With older boards
(pre-Diecimila) that lack auto-reset, you'll need to press the reset button on the board just
before starting the upload. On most boards, you'll see the RX and TX LEDs blink as the
sketch is uploaded. The Arduino Software (IDE) will display a message when the upload is
When you upload a sketch, you're using the Arduino bootloader, a small program that has
been loaded on to the microcontroller on your board. It allows you to upload code without
using any additional hardware. The bootloader is active for a few seconds when the board
resets; then it starts whichever sketch was most recently uploaded to the microcontroller. The
bootloader will blink the on-board (pin 13) LED when it starts (i.e., when the board resets).
Libraries
Libraries provide extra functionality for use in sketches, e.g., working with hardware or
manipulating data. To use a library in a sketch, select it from the Sketch > Import
Library menu. This will insert one or more #include statements at the top of the sketch and
compile the library with your sketch. Because libraries are uploaded to the board with your
sketch, they increase the amount of space it takes up. If a sketch no longer needs a library,
There is a list of libraries in the reference. Some libraries are included with the Arduino
software. Others can be downloaded from a variety of sources or through the Library
Manager. Starting with version 1.0.5 of the IDE, you do can import a library from a zip file
and use it in an open sketch. See these instructions for installing a third-party library.
directory. Platforms installed there may include board definitions (which appear in the board
the hardware directory, then unzip the third-party platform into its own sub-directory. (Don't
use "arduino" as the sub-directory name or you'll override the built-in Arduino platform.) To
For details on creating packages for third-party hardware, see the Arduino IDE 1.5 3rd party
Hardware specification.
Serial Monitor
This displays serial sent from the Arduino or Genuino board over USB or serial connector.
To send data to the board, enter text and click on the "send" button or press enter. Choose the
baud rate from the drop-down menu that matches the rate passed to Serial. Begin in your
sketch. Note that on Windows, Mac or Linux the board will reset (it will rerun your sketch)
when you connect with the serial monitor. Please note that the Serial Monitor does not
process control characters; if your sketch needs a complete management of the serial
communication with control characters, you can use an external terminal program and
page for details).
Preferences
Some preferences can be set in the preferences dialog (found under the Arduino menu on the
Mac, or File on Windows and Linux). The rest can be found in the preferences file, whose
Language Support
Since version 1.0.1 , the Arduino Software (IDE) has been translated into 30+ different
languages. By default, the IDE loads in the language selected by your operating system.
(Note: on Windows and possibly Linux, this is determined by the locale setting which
controls currency and date formats, not by the language the operating system is displayed in.)
If you would like to change the language manually, start the Arduino Software (IDE) and
currently supported languages. Select your preferred language from the menu, and restart the
software to use the selected language. If your operating system language is not supported, the
6 PWM.
Arduino Diecimila or Duemilanove w/ ATmega168
and 15 PWM.
LilyPad Arduino
About Proteus
ISIS is the software used to draw schematics and simulate the circuits in real time.
The simulation allows human access during run time, thus providing real time
simulation.
ARES is used for PCB designing. It has the feature of viewing output in 3D view of
Using PHP
With PHP, it’s a simple matter to embed dynamic activity in web pages. When you give
pages the .php extension, they have instant access to the scripting language. From a
developer’s point of view, all you have to do is write code such as the following: How are
you? The opening command. Outside of this construct, everything is sent to the client as
direct HTML. So, the text “How are you?” is simply output to the browser;
within the PHP tags, the built-in date function displays the current day of the week according
to the server’s system time. The final output of the two parts looks like this: Hello World.
Today is Wednesday. How are you? PHP is a flexible language, and some people prefer to
place the PHP construct directly next to PHP code, like this: Hello World. Today is . How are
you? There are also other ways of formatting and outputting information, which I’ll explain
in the chapters on PHP. The point is that with PHP, web developers have a scripting language
that, although not as fast as compiling your code in C or a similar language, is incredibly
Using PHP, you have unlimited control over your web server. Whether you need to modify
HTML on the fly, process a credit card, add user details to a database, or fetch information
from a third-party website, you can do it all from within the same PHP files in which the
Using MySQL
Of course, there’s not a lot of point to being able to change HTML output dynamically
unless you also have a means to track the changes that users make as they use your website.
In the early days of the Web, many sites used “flat” text files to store data such as usernames
and passwords. But this approach could cause problems if the file wasn’t correctly locked
against corruption from multiple simultaneous accesses. Also, a flat file can get only so big
before it becomes unwieldy to manage—not to mention the difficulty of trying to merge files
and perform complex searches in any kind of reasonable time. That’s where relational
databases with structured querying become essential. And MySQL, being free to use and
installed on vast numbers of Internet web servers, rises superbly to the occasion. It is a robust
and exceptionally fast database management system that uses English-like commands. The
highest level of MySQL structure is a database, within which you can have one or more
tables that contain your data. For example, let’s suppose you are working on a table called
users, within which you have created columns for surname, first name, and email, and you
now wish to add another user. One command that you might use to do this is: INSERT INTO
will have issued other commands to create the database and table and to set up all the correct
fields, but the INSERT command here shows how simple it can be to add new data to a
database. The INSERT command is an example of SQL (which stands for Structured Query
Language), a language designed in the early 1970s and reminiscent of one of the oldest
have an email address for a user and you need to look up that person’s name. To do this, you
In addition to PHP, MySQL, JavaScript, and CSS, there’s actually a fifth hero in the dynamic
Web: the web server. In the case of this book, that means the Apache web server. We’ve
discussed a little of what a web server does during the HTTP server/client exchange, but it
actually does much more behind the scenes. For example, Apache doesn’t serve up just
HTML files—it handles a wide range of files, from images and Flash files to MP3 audio files,
RSS (Really Simple Syndication) feeds, and more. Each element a web client encounters in
an HTML page is also requested from the server, which then serves it up. But these objects
don’t have to be static files, such as GIF images. They can all be generated by programs such
as PHP scripts. That’s right: PHP can even create images and other files for you, either on the
fly or in advance to serve up later. To do this, you normally have modules either precompiled
into Apache or PHP or called up at runtime. One such module is the GD library (short for
Apache also supports a huge range of modules of its own. In addition to the PHP module, the.
What Is a WAMP, MAMP, or LAMP?
WAMP, MAMP, and LAMP are abbreviations for “Windows, Apache, MySQL, and PHP,”
“Mac, Apache, MySQL, and PHP,” and “Linux, Apache, MySQL, and PHP,” 13 www.it-
ebooks.info respectively. These abbreviations describe a fully functioning setup used for
developing dynamic Internet web pages. WAMPs, MAMPs, and LAMPs come in the form of
a package that binds the bundled programs together so that you don’t have to install and set
them up separately. This means you can simply download and install a single program and
follow a few easy prompts to get your web development server up and running in the quickest
Accuracy
Accuracy is an evaluation metric that is used for measuring the performance of classification
or regression algorithms. Accuracy can be a problematic or misleading performance metric
when used to evaluate a model that is trained on unbalanced data. For this evaluation metric
to provide a good and reliable performance measure, the datato be used in training the model
must be balanced. Accuracy is computed by summing up the true positive (TP) and true
negative (TN) classes divided by the summation of true positive, true negative, false positive
(FP), and false negative (FN) classes, as shown in the formula below.
Accuracy =TP + TN / TP + TN + FP + FN
Recall
Recall is another evaluation metric that is used to measure the performance of the classifier.
Recall is a correctly classified class from the classification model. Recall is computed by
dividing the true positive class by the sum of the true positive class and false-negative class,
as shown in the formula below.
Recall = TP / TP + FN
Precision
Precision is also an evaluation metric usually used together with the recall evaluation metric
to measure the performance of classification algorithms. Precision is a positive prediction
result that corresponds to the correctly classified class. Precision is computed by dividing the
true positive class by the sum of the true positive and false positive classes, as shown in the
formula below.
Precision = TP / TP + FP
F1 Score
F1 score evaluation metric is used to measure the performance of a classifier by combining
the recall andprecision evaluation metrics as one single performance measure evaluation
metric. F1 score evaluation metric is computed by multiplying the product of the precision
and recall by 2 and dividing by the sum of recall and precision evaluation metrics. 'e formula
below shows how the F1 score evaluation metric is computed.
F1 Score = 2 × (precision × recall)/precision + recall
CHAPTER 8
CONCLUSION AND FUTURE ENHANCEMENT
8.1. Conclusion
The recent Asthma has exposed the world to very serious challenges in fast diagnostics and
monitoring of the outbreak. Selective sensing approaches that rely on specific and well-
defined targets, such as in PCR, have been adopted toward fast diagnostics, but substantial
pitfalls still exist. Indeed, such detection techniques are very disease-specific and their
adaptation in the case of SARS-CoV-2 mutations requires significant effort and time. On the
other hand, the use of a nonspecific sensing approach, mainly using breath samples, could go
a long way toward healthful, responsible self-care.The e-Nose system was successfully
designedand implemented to classify the exhaled breath. The methodaccuracies were 92%
Support Vector Machine—94.87%.
8.2. Future Enhancement
In future, E-Nose can be used as a platform for other medical conditions as well as meeting
its original goal of monitoring astronaut health and crew cabin air quality.
REFERENCES
1. V. Guarrasi, N. C. D'Amico, R. Sicilia, E. Cordelli, and P. Soda,``Pareto optimization
of deep networks for ASTHMA DISEASEdiagnosis fromchest X-rays,'' Pattern
Recognit., vol. 121, Jan. 2022, Art. no. 108242.[Online]. Available:
https://www.sciencedirect.com/science/article/pii/S0031320321004234
2. M. Kaur, V. Kumar, V. Yadav, D. Singh, N. Kumar, and N. N. Das, “Metaheuristic-
based deep ASTHMA DISEASEscreening model from chest X-ray images,” Journal
of Healthcare Engineering, vol. 2020, Article ID 8829829, 17 pages, 2021.
3. P. Soda et al., ``AIforCOVID: Predicting the clinical outcomes inpatients with
ASTHMA DISEASEapplying AI to chest-X-rays. An Italianmulticentre study,'' Med.
Image Anal., vol. 74, Dec. 2021,Art. no. 102216. [Online]. Available:
https://www.sciencedirect.com/science/article/pii/S1361841521002619
4. O. D. T. Catala, I. S. Igual, F. J. Perez-Benito, D. M. Escriva, V. O. Castello,R.
Llobet, and J.-C. Perez-Cortes, ``Bias analysis on public X-ray imagedatasets of
pneumonia and ASTHMA DISEASEpatients,'' IEEE Access, vol. 9, pp. 42370-42383,
2021.
5. Y. Bouchareb, P. M. Khaniabadi, F. Al Kindi, H. Al Dhuhli,I. Shiri, H. Zaidi, and A.
Rahmim, ``Artificial intelligence-drivenassessment of radiological images for
ASTHMA DISEASE,'' Comput. Biol.Med., vol. 136, Sep. 2021, Art. no. 104665.
[Online].
Available:https://www.sciencedirect.com/science/article/pii/S0010482521004595
6. D. Apostolopoulos and T. A. Mpesiana, “Asthma Disease: Automatic detection from
X-ray images utilizing transfer learning with convolutional neural networks,” Phys.
Eng. Sci. Med., vol. 43, no. 2, pp. 635–640, Jun. 2020.
7. A. Sharma, S. Rani, and D. Gupta, “Artificial intelligencebasedclassification of chest
X-ray images into ASTHMA DISEASEandother infectious diseases,” International
Journal of BiomedicalImaging, vol. 2020, Article ID 8889023, 5 pages, 2020.
8. F. Ucar and D. Korkmaz, “COVIDiagnosis-Net: deep bayessqueezenetbased
diagnosis of the coronavirus disease 2019(ASTHMA DISEASE) from X-ray images,”
Medical Hypotheses,vol. 140, Article ID 109761, 2020.
9. Y. Fang et al., “Sensitivity of chest CT for ASTHMA DISEASE: Comparison toRT-
PCR,” Radiology, vol. 296, no. 2, pp. E115–E117, Aug. 2020.
10. M. Shorfuzzaman and M. Masud, “On the detection of Asthma Diseasefrom chest x-
ray images using cnn-based transfer learning,”Computers, Materials & Continua, vol.
64, no. 3,pp. 1359–1381, 2020.
11. W. Xie, C. Jacobs, J.-P. Charbonnier, and B. van Ginneken, “Relationalmodeling for
robust and efficient pulmonary lobe segmentation inCT scans,” IEEE Trans. Med.
Imag., vol. 39, no. 8, pp. 2664–2675,Aug. 2020.
12. J. Irvin, P. Rajpurkar, M.Ko,Y.Yu, S. Ciurea-Ilcus, C. Chute, H. Marklund,B.
Haghgoo, R. Ball, K. Shpanskaya, J. Seekins, D. A. Mong, S. S. Halabi,J. K.
Sandberg, R. Jones, D. B. Larson, C. P. Langlotz, B. N. Patel,M. P. Lungren, and A.
Y. Ng, ``CheXpert: A large chest radiograph datasetwith uncertainty labels and expert
comparison,'' in Proc. AAAI Conf. Artif.Intell., 2019, vol. 33, no. 1, pp. 590-597.
13. G. S. Collins and K. G. M. Moons, ``Reporting of artiticial intelligence
predictionmodels,'' Lancet, vol. 393, no. 10181, pp. 1577-1579, Apr. 2019.
14. F. Yan, X. Huang, Y. Yao, M. Lu, and M. Li, ``CombiningLSTM and DenseNet for
automatic annotation and classificationof chest X-ray images,'' IEEE Access, vol. 7,
pp. 74181-74189,2019.
15. J. Irvin, P. Rajpurkar, M. Ko, and Y. Yu, ``Chexpert: A large chest radiographdataset
with uncertainty labels and expert comparison,'' in Proc.AAAI Conf. Artif. Intell., vol.
33, 2019, pp. 590-597.