DNN Architecture for Waste Segregation
This research aims at using Fuzzy Inference Systems (FIS) to evaluate the workloads which Hadoop intakes and passes that information to the capacity manager. We aim at checking the performance of FIS as compared to various techniques for data distribution.
23 Pages6469 Words246 Views
Added on 2022-10-15
About This Document
This proposal discusses how automated systems based on Convolutional Neural Network architectures can improve food waste segregation. It provides a background of the study, problem statement, rationale, aims and objectives, research questions, and literature review. The study aims to create an automated system that detects and identifies food waste with the help of a deep neural networking-based architecture.
DNN Architecture for Waste Segregation
This research aims at using Fuzzy Inference Systems (FIS) to evaluate the workloads which Hadoop intakes and passes that information to the capacity manager. We aim at checking the performance of FIS as compared to various techniques for data distribution.
Added on 2022-10-15
ShareRelated Documents
Running head: DNN ARCHITECTURE FOR WASTE SEGREGATION
DNN Architecture for Waste Segregation
Name of the Student
Name of the University
Author Note
DNN Architecture for Waste Segregation
Name of the Student
Name of the University
Author Note
DNN ARCHITECTURE FOR WASTE SEGREGATION
1
Abstract
According to the paper ‘Food detection and recognition using convolutional
neural networking,’ it can be understood how food waste segregation can be
improved through Convolutional Neural Network architectures. The report first
provides a background of the study where it mentions the menace of food
wastage and how it is affecting the society, the present-day labourers who deal
with the wastes, the community as well as those suffering from malnutrition in
the underdeveloped countries and also notes how this can disturb the wellbeing
of people as also give rise to newer diseases. Then the study presents the
problem statement where it discusses in detail the negative aspects of treating
wastes through manual labourers along with highlighting the three key use cases
of recycling the food waste like compost for feeding animals, conversion of food
waste to biogas for energy and reuse of food packaging. Then the study jumps to
the rationale where it describes how the automated system using CNN can
address the problems highlighted in the earlier sections. After this the study
provides the aims and objectives of the research which are identified to be the
viability of creating the automated system, the types of classification that the
system should make, the deep learning architecture suitable for the system as
also how the architecture can help the system. Then after presenting the
research questions the study enters into the literature review section where the
study talks about the advantages of artificial neural networks (ANN) and why
CNN can help better in segregating food waste from regular waste. Next the
study explains the methodology of the CNN architecture and how the deep
learning solution helps the automated system meet the objectives after which
the proposal ends with observations in concluding notes.
1
Abstract
According to the paper ‘Food detection and recognition using convolutional
neural networking,’ it can be understood how food waste segregation can be
improved through Convolutional Neural Network architectures. The report first
provides a background of the study where it mentions the menace of food
wastage and how it is affecting the society, the present-day labourers who deal
with the wastes, the community as well as those suffering from malnutrition in
the underdeveloped countries and also notes how this can disturb the wellbeing
of people as also give rise to newer diseases. Then the study presents the
problem statement where it discusses in detail the negative aspects of treating
wastes through manual labourers along with highlighting the three key use cases
of recycling the food waste like compost for feeding animals, conversion of food
waste to biogas for energy and reuse of food packaging. Then the study jumps to
the rationale where it describes how the automated system using CNN can
address the problems highlighted in the earlier sections. After this the study
provides the aims and objectives of the research which are identified to be the
viability of creating the automated system, the types of classification that the
system should make, the deep learning architecture suitable for the system as
also how the architecture can help the system. Then after presenting the
research questions the study enters into the literature review section where the
study talks about the advantages of artificial neural networks (ANN) and why
CNN can help better in segregating food waste from regular waste. Next the
study explains the methodology of the CNN architecture and how the deep
learning solution helps the automated system meet the objectives after which
the proposal ends with observations in concluding notes.
DNN ARCHITECTURE FOR WASTE SEGREGATION
2
Contents
1 Introduction..............................................................................................3
2 Proposal....................................................................................................3
2.1 Background of the Study....................................................................3
2.2 Problem Statement............................................................................4
2.3 Rational of Research..........................................................................4
2.4 Research aims and objectives............................................................5
2.5 Research questions............................................................................5
2.6 Research hypothesis..........................................................................5
2.7 Literature review................................................................................6
2.7.1 Use of Hidden layers....................................................................7
2.7.2 Extraction of image to Hidden Layers..........................................7
2.7.3 Use of Multiple Hidden Layers......................................................8
2.7.4 RileyKwok Food Classification......................................................8
2.7.5 Supervised Food classification using DCNN.................................9
2.8 Research Methods and Specification..................................................9
2.8.1 CNN Architecture Design.............................................................9
2.8.2 Confusion Matrix for CNN model................................................10
2.8.3 Process Flow of Automated System...........................................11
2.8.4 Implementation and Deployment..............................................12
2.8.5 Software used for training the system.......................................12
2.8.6 Research Findings......................................................................13
2.8.7 Timeline of the study.................................................................14
3 References.............................................................................................15
2
Contents
1 Introduction..............................................................................................3
2 Proposal....................................................................................................3
2.1 Background of the Study....................................................................3
2.2 Problem Statement............................................................................4
2.3 Rational of Research..........................................................................4
2.4 Research aims and objectives............................................................5
2.5 Research questions............................................................................5
2.6 Research hypothesis..........................................................................5
2.7 Literature review................................................................................6
2.7.1 Use of Hidden layers....................................................................7
2.7.2 Extraction of image to Hidden Layers..........................................7
2.7.3 Use of Multiple Hidden Layers......................................................8
2.7.4 RileyKwok Food Classification......................................................8
2.7.5 Supervised Food classification using DCNN.................................9
2.8 Research Methods and Specification..................................................9
2.8.1 CNN Architecture Design.............................................................9
2.8.2 Confusion Matrix for CNN model................................................10
2.8.3 Process Flow of Automated System...........................................11
2.8.4 Implementation and Deployment..............................................12
2.8.5 Software used for training the system.......................................12
2.8.6 Research Findings......................................................................13
2.8.7 Timeline of the study.................................................................14
3 References.............................................................................................15
DNN ARCHITECTURE FOR WASTE SEGREGATION
3
1 Introduction
The following study attempts to explain how food waste segregation can
be improved through automated systems based on Convolutional Neural
Network architectures (Jagtap et al. 2019). The report starts by providing a
background of the study where it mentions the menace of food wastage and how
it is affecting the society, the present-day labourers who deal with the wastes,
the community as well as those suffering from malnutrition in the
underdeveloped countries and also notes how this can disturb the wellbeing of
people as also give rise to newer diseases (Rucevska et al. 2017). Then the study
presents the problem statement where it discusses in detail the negative aspects
of treating wastes through manual labourers along with highlighting the three
key use cases of recycling the food waste like compost for feeding animals,
conversion of food waste to biogas for energy and reuse of food packaging. Then
the study jumps to the rationale where it describes how the automated system
using CNN can address the problems highlighted in the earlier sections (Van Phat
et al. 2017). After this the study provides the aims and objectives of the research
which are identified to be the viability of creating the automated system, the
types of classification that the system should make, the deep learning
architecture suitable for the system as also how the architecture can help the
system. Then after presenting the research questions the study enters into the
literature review section where the study talks about the advantages of artificial
neural networks (ANN) and why CNN can help better in segregating food waste
from regular waste. Next the study explains the methodology of the CNN
architecture and how the deep learning solution helps the automated system
meet the objectives after which the proposal ends with observations in
concluding notes.
2 Proposal
2.1 Background of the Study
Treatment of leftover food is increasingly growing as a concerning factor
in the society. Several reports suggest that about 58% of the total food produced
by individuals are getting wasted all around the world on a daily basis (Aizawa et
al. 2014). This compound the ongoing problem of 60% of the human population
in the third world countries are struggling to find food for sustaining their survival
process. This is resulting in increased number of people suffering from
malnutrition. These people are thus resorting to scavenging of food and thus
3
1 Introduction
The following study attempts to explain how food waste segregation can
be improved through automated systems based on Convolutional Neural
Network architectures (Jagtap et al. 2019). The report starts by providing a
background of the study where it mentions the menace of food wastage and how
it is affecting the society, the present-day labourers who deal with the wastes,
the community as well as those suffering from malnutrition in the
underdeveloped countries and also notes how this can disturb the wellbeing of
people as also give rise to newer diseases (Rucevska et al. 2017). Then the study
presents the problem statement where it discusses in detail the negative aspects
of treating wastes through manual labourers along with highlighting the three
key use cases of recycling the food waste like compost for feeding animals,
conversion of food waste to biogas for energy and reuse of food packaging. Then
the study jumps to the rationale where it describes how the automated system
using CNN can address the problems highlighted in the earlier sections (Van Phat
et al. 2017). After this the study provides the aims and objectives of the research
which are identified to be the viability of creating the automated system, the
types of classification that the system should make, the deep learning
architecture suitable for the system as also how the architecture can help the
system. Then after presenting the research questions the study enters into the
literature review section where the study talks about the advantages of artificial
neural networks (ANN) and why CNN can help better in segregating food waste
from regular waste. Next the study explains the methodology of the CNN
architecture and how the deep learning solution helps the automated system
meet the objectives after which the proposal ends with observations in
concluding notes.
2 Proposal
2.1 Background of the Study
Treatment of leftover food is increasingly growing as a concerning factor
in the society. Several reports suggest that about 58% of the total food produced
by individuals are getting wasted all around the world on a daily basis (Aizawa et
al. 2014). This compound the ongoing problem of 60% of the human population
in the third world countries are struggling to find food for sustaining their survival
process. This is resulting in increased number of people suffering from
malnutrition. These people are thus resorting to scavenging of food and thus
DNN ARCHITECTURE FOR WASTE SEGREGATION
4
causing rise of newer unknown diseases in the society (Holloway 2017). Manual
labour tasked with recycling the waste is also vulnerable to several diseases and
are unable lead a healthy life. As a result, it is understood that automating the
waste segregating process along with identification and recognition of the food
waste can help address the problems plaguing the society and help in increasing
the living standards of the associated people.
2.2 Problem Statement
Segregation of food waste and subsequent recycling of these waste
materials is an important issue for sustaining balanced standards of living in the
society. The processes of current day involve manual labour where people are
tasked with manually differentiating food waste from waste materials and feed
more granular waste to filters for further segregation of waste. This is affecting
the society as the workers are paid very less and the work ends up negatively
affecting their wellbeing by exposing them to risk of inflicting diseases from
various viruses and bacteria as well as affecting their health due to excess work
hours. Work of these labourers thus need to be transformed to empower them
with a better work for a better life (Zhou, Dang-Nguyen and Gurrin 2017). If the
waste segregation can be made accurate and efficient, it can enable the society
in several ways. These can be donating of the recycled food for animal feeding,
converting the food waste to obtain biogas and thus generate energy where it is
not readily available as also the packaging material for food can be reused for
different purposes. Additionally, education programs can be created by assessing
the data from segregation of food to spread awareness among people of the
society in ensuring that the least amount of food is wasted in future so that more
food can be made available for the needy in the society or allocated as aids for
poorer countries (Ishak, Romli and Rahman 2018). Hence an automated system
has to be introduced that can accurately detect and identify food from other
waste and is also able to identify what is type of waste are to be used in the
filters.
2.3 Rational of Research
The increase in regular wastage of food is affecting the people of the
society in negative way. To improve the current process of managing waste and
recycling of food waste through labourers, an automated system needs to be
introduced. For this automated system to be accurate and efficient in detecting
food and identifying the contents, machine learning techniques need to be used
4
causing rise of newer unknown diseases in the society (Holloway 2017). Manual
labour tasked with recycling the waste is also vulnerable to several diseases and
are unable lead a healthy life. As a result, it is understood that automating the
waste segregating process along with identification and recognition of the food
waste can help address the problems plaguing the society and help in increasing
the living standards of the associated people.
2.2 Problem Statement
Segregation of food waste and subsequent recycling of these waste
materials is an important issue for sustaining balanced standards of living in the
society. The processes of current day involve manual labour where people are
tasked with manually differentiating food waste from waste materials and feed
more granular waste to filters for further segregation of waste. This is affecting
the society as the workers are paid very less and the work ends up negatively
affecting their wellbeing by exposing them to risk of inflicting diseases from
various viruses and bacteria as well as affecting their health due to excess work
hours. Work of these labourers thus need to be transformed to empower them
with a better work for a better life (Zhou, Dang-Nguyen and Gurrin 2017). If the
waste segregation can be made accurate and efficient, it can enable the society
in several ways. These can be donating of the recycled food for animal feeding,
converting the food waste to obtain biogas and thus generate energy where it is
not readily available as also the packaging material for food can be reused for
different purposes. Additionally, education programs can be created by assessing
the data from segregation of food to spread awareness among people of the
society in ensuring that the least amount of food is wasted in future so that more
food can be made available for the needy in the society or allocated as aids for
poorer countries (Ishak, Romli and Rahman 2018). Hence an automated system
has to be introduced that can accurately detect and identify food from other
waste and is also able to identify what is type of waste are to be used in the
filters.
2.3 Rational of Research
The increase in regular wastage of food is affecting the people of the
society in negative way. To improve the current process of managing waste and
recycling of food waste through labourers, an automated system needs to be
introduced. For this automated system to be accurate and efficient in detecting
food and identifying the contents, machine learning techniques need to be used
DNN ARCHITECTURE FOR WASTE SEGREGATION
5
(Ramalingam et al. 2018). Machine learning includes a wide range of
technologies and algorithms in making a system learn to perform a task and
hence the appropriate neural networking architecture and corresponding
algorithms needs to be used so that the deep learning solution is ideal for the
problems it aims to address (Yoshiyuki and Keiji 2014). The technologies and
techniques to be used by this neural networking architecture needs are to ensure
that they enable the system in further increasing the accuracy of the generated
output such that not only can human personnel can be freed from the manual
task of treating wastes, but more accurate and consistent results can be
obtained as well as enabling societies in spreading awareness among the people
on preventing further wastage of food so that more aid can be arranged for
poorer people in underdeveloped countries.
2.4 Research aims and objectives
The aim of this study is about understanding how the involvement of
manual labour in treating and recycling food waste from regular waste can be
eliminated by introducing an automated system that detects and identifies food
waste with the help of a deep neural networking-based architecture.
The aims and objectives of this study is the following:
i. To understand the means of creating an automated system using deep
neural networking architecture to distinguish food waste from regular
waste for subsequent recycling.
ii. To identify the nutritional value of the wasted food and thus help in
educating the community on usefulness of these foods.
iii. To identify the quantity of wasted food such that awareness programs be
created for the people to help reduce wastage of food
iv. To research about machine learning technologies that can help create the
automated solution such that the problems can be addressed
v. To analyse how the deep learning architecture is used by the system to
accurately classify food waste for performing the tasks assigned to it
2.5 Research questions
Question 1: What makes it possible to implement the automated system using
the required neural networking architecture for identifying food waste for
recycling?
Question 2: Can the automated system identify nutritional value of wasted food?
5
(Ramalingam et al. 2018). Machine learning includes a wide range of
technologies and algorithms in making a system learn to perform a task and
hence the appropriate neural networking architecture and corresponding
algorithms needs to be used so that the deep learning solution is ideal for the
problems it aims to address (Yoshiyuki and Keiji 2014). The technologies and
techniques to be used by this neural networking architecture needs are to ensure
that they enable the system in further increasing the accuracy of the generated
output such that not only can human personnel can be freed from the manual
task of treating wastes, but more accurate and consistent results can be
obtained as well as enabling societies in spreading awareness among the people
on preventing further wastage of food so that more aid can be arranged for
poorer people in underdeveloped countries.
2.4 Research aims and objectives
The aim of this study is about understanding how the involvement of
manual labour in treating and recycling food waste from regular waste can be
eliminated by introducing an automated system that detects and identifies food
waste with the help of a deep neural networking-based architecture.
The aims and objectives of this study is the following:
i. To understand the means of creating an automated system using deep
neural networking architecture to distinguish food waste from regular
waste for subsequent recycling.
ii. To identify the nutritional value of the wasted food and thus help in
educating the community on usefulness of these foods.
iii. To identify the quantity of wasted food such that awareness programs be
created for the people to help reduce wastage of food
iv. To research about machine learning technologies that can help create the
automated solution such that the problems can be addressed
v. To analyse how the deep learning architecture is used by the system to
accurately classify food waste for performing the tasks assigned to it
2.5 Research questions
Question 1: What makes it possible to implement the automated system using
the required neural networking architecture for identifying food waste for
recycling?
Question 2: Can the automated system identify nutritional value of wasted food?
End of preview
Want to access all the pages? Upload your documents or become a member.
Related Documents
A Presentation on Waste Segregation using DNN.lg...
|8
|491
|3
Convolution Neural Networks Research 2022lg...
|10
|2688
|35
A Robust Deep Neural Network Based Breast Cancer Detection And Classification Abstract: Case Centric Segmentation, Feature Extraction and Classificationlg...
|16
|14251
|98
Paired Facial Matching With Age and Gender Prediction | Reportlg...
|8
|2417
|9