No Cover Image

E-Thesis 150 views

Irregular Domain Deep Learning / SACHIN BAHADE

Swansea University Author: SACHIN BAHADE

DOI (Published version): 10.23889/SUthesis.65003

Abstract

In recent years, the use of machine learning and deep learning on graph data has increased significantly. Convolutional neural networks have achieved remarkable success with grid-like data such as images, but encounter enormous difficulties when learning from more general structures such as graphs....

Full description

Published: Swansea, Wales, UK 2023
Institution: Swansea University
Degree level: Doctoral
Degree name: Ph.D
Supervisor: Xie, Xianghua. and Edwards, Mike.
URI: https://cronfa.swan.ac.uk/Record/cronfa65003
first_indexed 2023-11-17T11:37:19Z
last_indexed 2024-11-25T14:15:10Z
id cronfa65003
recordtype RisThesis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2024-11-04T10:39:06.4437764</datestamp><bib-version>v2</bib-version><id>65003</id><entry>2023-11-17</entry><title>Irregular Domain Deep Learning</title><swanseaauthors><author><sid>a4bcef8d532a3fd7bba3e7c6c0471779</sid><firstname>SACHIN</firstname><surname>BAHADE</surname><name>SACHIN BAHADE</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2023-11-17</date><abstract>In recent years, the use of machine learning and deep learning on graph data has increased significantly. Convolutional neural networks have achieved remarkable success with grid-like data such as images, but encounter enormous difficulties when learning from more general structures such as graphs. The inclusion of trainable local filters enables the automated extraction of high-level features in a regular domain; however, due to the irregular structure of graph data, regular domain deep learning operations are constrained. Thus, the development of deep learning models that can successfully learn from graph data is a promising research field with a great potential for impact. In this work, we are looking to generalise deep learning model for learning spatial feature representations on irregular domain topologies for the purpose of learning high-level model (structures, topology, parts) in medical image analysis problem. This thesis focuses specifically on graph neural networks as the primary machine learning model for effectively tacking graph data challenges. This study investigates the use of irregular domains deep learning methods to enhance the high-level model, with applications in medical image segmentation and detection tasks. We first explore how graph construction affects the behaviour of graph convolutional operations. For this purpose, we use a parametrically specified graph to represent a localised sampling operation on an underlying domain, which we subsequently mine for features while analysing the effect of the graph&#x2019;s construction on the model&#x2019;s behaviour. After this, we learn features using advanced deep-learning, spectral, and spatial-based graph signal processing techniques for cell segmentation on immunostained images and present a comparative study. Third, we explore the problem of object recognition in an irregular environment. Conventional convolutional neural networks may process Euclidean data for object detection tasks in a number of ways, but the use of graphs to sample from image data requires special consideration, which has the possibility of generalising to non-Euclidean data. We describe a graph convolution based region proposal method for the detection of non-Euclidean data objects. The extraction of a subgraph as a candidate for the prospective object regions is our primary focus. We discovered improvements when comparing our technique to the region based convolutional neural networks method for the Euclidean domain. The last main chapter focuses on the problem of nuclei detection and finds that graph convolutional networks-based cascaded architecture outperforms convolutional neural networks-based techniques and is more stable. In conclusion, we demonstrate the stability of our irregular domain deep learning methods for graph construction, cell segmentation, and nucleus detection applications, as well as its improved performance in comparison to convolutional neural networks-based approaches.</abstract><type>E-Thesis</type><journal/><volume/><journalNumber/><paginationStart/><paginationEnd/><publisher/><placeOfPublication>Swansea, Wales, UK</placeOfPublication><isbnPrint/><isbnElectronic/><issnPrint/><issnElectronic/><keywords>Graph Neural Network, Biomedical Imaging, Machine Learning, Deep Learning</keywords><publishedDay>26</publishedDay><publishedMonth>10</publishedMonth><publishedYear>2023</publishedYear><publishedDate>2023-10-26</publishedDate><doi>10.23889/SUthesis.65003</doi><url/><notes/><college>COLLEGE NANME</college><CollegeCode>COLLEGE CODE</CollegeCode><institution>Swansea University</institution><supervisor>Xie, Xianghua. and Edwards, Mike.</supervisor><degreelevel>Doctoral</degreelevel><degreename>Ph.D</degreename><degreesponsorsfunders>National Overseas Scholarship Scheme for Scheduled Castes, Ministry of Social Justice and Empowerment, India.</degreesponsorsfunders><apcterm/><funders>National Overseas Scholarship Scheme for Scheduled Castes, Ministry of Social Justice and Empowerment, India.</funders><projectreference/><lastEdited>2024-11-04T10:39:06.4437764</lastEdited><Created>2023-11-17T11:32:32.2065407</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Mathematics and Computer Science - Computer Science</level></path><authors><author><firstname>SACHIN</firstname><surname>BAHADE</surname><order>1</order></author></authors><documents><document><filename>65003__29044__29e5004d0a2344c8b04aeb9c9b764b92.pdf</filename><originalFilename>2023_Bahade_SS.final.65003.pdf</originalFilename><uploaded>2023-11-17T11:37:44.5352053</uploaded><type>Output</type><contentLength>13321459</contentLength><contentType>application/pdf</contentType><version>E-Thesis</version><cronfaStatus>true</cronfaStatus><embargoDate>2024-10-26T00:00:00.0000000</embargoDate><documentNotes>Copyright: The Author, Sachin S. Bahade, 2023.</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language></document></documents><OutputDurs/></rfc1807>
spelling 2024-11-04T10:39:06.4437764 v2 65003 2023-11-17 Irregular Domain Deep Learning a4bcef8d532a3fd7bba3e7c6c0471779 SACHIN BAHADE SACHIN BAHADE true false 2023-11-17 In recent years, the use of machine learning and deep learning on graph data has increased significantly. Convolutional neural networks have achieved remarkable success with grid-like data such as images, but encounter enormous difficulties when learning from more general structures such as graphs. The inclusion of trainable local filters enables the automated extraction of high-level features in a regular domain; however, due to the irregular structure of graph data, regular domain deep learning operations are constrained. Thus, the development of deep learning models that can successfully learn from graph data is a promising research field with a great potential for impact. In this work, we are looking to generalise deep learning model for learning spatial feature representations on irregular domain topologies for the purpose of learning high-level model (structures, topology, parts) in medical image analysis problem. This thesis focuses specifically on graph neural networks as the primary machine learning model for effectively tacking graph data challenges. This study investigates the use of irregular domains deep learning methods to enhance the high-level model, with applications in medical image segmentation and detection tasks. We first explore how graph construction affects the behaviour of graph convolutional operations. For this purpose, we use a parametrically specified graph to represent a localised sampling operation on an underlying domain, which we subsequently mine for features while analysing the effect of the graph’s construction on the model’s behaviour. After this, we learn features using advanced deep-learning, spectral, and spatial-based graph signal processing techniques for cell segmentation on immunostained images and present a comparative study. Third, we explore the problem of object recognition in an irregular environment. Conventional convolutional neural networks may process Euclidean data for object detection tasks in a number of ways, but the use of graphs to sample from image data requires special consideration, which has the possibility of generalising to non-Euclidean data. We describe a graph convolution based region proposal method for the detection of non-Euclidean data objects. The extraction of a subgraph as a candidate for the prospective object regions is our primary focus. We discovered improvements when comparing our technique to the region based convolutional neural networks method for the Euclidean domain. The last main chapter focuses on the problem of nuclei detection and finds that graph convolutional networks-based cascaded architecture outperforms convolutional neural networks-based techniques and is more stable. In conclusion, we demonstrate the stability of our irregular domain deep learning methods for graph construction, cell segmentation, and nucleus detection applications, as well as its improved performance in comparison to convolutional neural networks-based approaches. E-Thesis Swansea, Wales, UK Graph Neural Network, Biomedical Imaging, Machine Learning, Deep Learning 26 10 2023 2023-10-26 10.23889/SUthesis.65003 COLLEGE NANME COLLEGE CODE Swansea University Xie, Xianghua. and Edwards, Mike. Doctoral Ph.D National Overseas Scholarship Scheme for Scheduled Castes, Ministry of Social Justice and Empowerment, India. National Overseas Scholarship Scheme for Scheduled Castes, Ministry of Social Justice and Empowerment, India. 2024-11-04T10:39:06.4437764 2023-11-17T11:32:32.2065407 Faculty of Science and Engineering School of Mathematics and Computer Science - Computer Science SACHIN BAHADE 1 65003__29044__29e5004d0a2344c8b04aeb9c9b764b92.pdf 2023_Bahade_SS.final.65003.pdf 2023-11-17T11:37:44.5352053 Output 13321459 application/pdf E-Thesis true 2024-10-26T00:00:00.0000000 Copyright: The Author, Sachin S. Bahade, 2023. true eng
title Irregular Domain Deep Learning
spellingShingle Irregular Domain Deep Learning
SACHIN BAHADE
title_short Irregular Domain Deep Learning
title_full Irregular Domain Deep Learning
title_fullStr Irregular Domain Deep Learning
title_full_unstemmed Irregular Domain Deep Learning
title_sort Irregular Domain Deep Learning
author_id_str_mv a4bcef8d532a3fd7bba3e7c6c0471779
author_id_fullname_str_mv a4bcef8d532a3fd7bba3e7c6c0471779_***_SACHIN BAHADE
author SACHIN BAHADE
author2 SACHIN BAHADE
format E-Thesis
publishDate 2023
institution Swansea University
doi_str_mv 10.23889/SUthesis.65003
college_str Faculty of Science and Engineering
hierarchytype
hierarchy_top_id facultyofscienceandengineering
hierarchy_top_title Faculty of Science and Engineering
hierarchy_parent_id facultyofscienceandengineering
hierarchy_parent_title Faculty of Science and Engineering
department_str School of Mathematics and Computer Science - Computer Science{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Mathematics and Computer Science - Computer Science
document_store_str 1
active_str 0
description In recent years, the use of machine learning and deep learning on graph data has increased significantly. Convolutional neural networks have achieved remarkable success with grid-like data such as images, but encounter enormous difficulties when learning from more general structures such as graphs. The inclusion of trainable local filters enables the automated extraction of high-level features in a regular domain; however, due to the irregular structure of graph data, regular domain deep learning operations are constrained. Thus, the development of deep learning models that can successfully learn from graph data is a promising research field with a great potential for impact. In this work, we are looking to generalise deep learning model for learning spatial feature representations on irregular domain topologies for the purpose of learning high-level model (structures, topology, parts) in medical image analysis problem. This thesis focuses specifically on graph neural networks as the primary machine learning model for effectively tacking graph data challenges. This study investigates the use of irregular domains deep learning methods to enhance the high-level model, with applications in medical image segmentation and detection tasks. We first explore how graph construction affects the behaviour of graph convolutional operations. For this purpose, we use a parametrically specified graph to represent a localised sampling operation on an underlying domain, which we subsequently mine for features while analysing the effect of the graph’s construction on the model’s behaviour. After this, we learn features using advanced deep-learning, spectral, and spatial-based graph signal processing techniques for cell segmentation on immunostained images and present a comparative study. Third, we explore the problem of object recognition in an irregular environment. Conventional convolutional neural networks may process Euclidean data for object detection tasks in a number of ways, but the use of graphs to sample from image data requires special consideration, which has the possibility of generalising to non-Euclidean data. We describe a graph convolution based region proposal method for the detection of non-Euclidean data objects. The extraction of a subgraph as a candidate for the prospective object regions is our primary focus. We discovered improvements when comparing our technique to the region based convolutional neural networks method for the Euclidean domain. The last main chapter focuses on the problem of nuclei detection and finds that graph convolutional networks-based cascaded architecture outperforms convolutional neural networks-based techniques and is more stable. In conclusion, we demonstrate the stability of our irregular domain deep learning methods for graph construction, cell segmentation, and nucleus detection applications, as well as its improved performance in comparison to convolutional neural networks-based approaches.
published_date 2023-10-26T14:29:11Z
_version_ 1821325500002336768
score 11.048042

Similar Items