\"The main focus of this work is to use machine learning and data mining techniques to address some challenging problems that arise from nuclear data. Specifically, two problem areas are discussed: nuclear imaging and radiation detection. The techniques to approach these problems are primarily based on a variant of Artificial Neural Network (ANN) called Convolutional Neural Network (CNN), which is one of the most popular forms of 'deep learning' technique.
The first problem is about interpreting and analyzing 3D medical radiation images automatically. A method is developed to identify and quantify deformable image registration (DIR) errors from lung CT scans for quality assurance (QA) purposes. The method includes the process of preparing a CT scan dataset for machine learning model training and testing, design of a 3D convolutional neural network architecture that classifies registrations into good or poor classes and a metric called Registration Error Index (REI), which gives us a quantitative measure of registration error. The method achieves 0.882 AUC-ROC on the test dataset. Furthermore, the combined standard uncertainty of the estimated REI by our model lies within 11%, with a confidence level of approximately 68%.
The second problem is about automatic radioisotope identification in real-time. A comparison of five different machine learning models is presented for gamma spectroscopy at a wide variety of testing conditions. Moreover, Hybrid Neural Network (HNN) was developed specifically for gamma-ray spectra data. Three different methods for feature extraction were tested as well. Experiments on MCNP simulated spectra suggest that HNN can achieve 2-12% higher F1 score at difficult testing conditions compared to best performing traditional ML models and obtains 93.33% F1 score during evaluation\"--Abstract, page iii.
","abstract_html":"<p>"The main focus of this work is to use machine learning and data mining techniques to address some challenging problems that arise from nuclear data. Specifically, two problem areas are discussed: nuclear imaging and radiation detection. The techniques to approach these problems are primarily based on a variant of Artificial Neural Network (ANN) called Convolutional Neural Network (CNN), which is one of the most popular forms of 'deep learning' technique.</p> <p>The first problem is about interpreting and analyzing 3D medical radiation images automatically. A method is developed to identify and quantify deformable image registration (DIR) errors from lung CT scans for quality assurance (QA) purposes. The method includes the process of preparing a CT scan dataset for machine learning model training and testing, design of a 3D convolutional neural network architecture that classifies registrations into good or poor classes and a metric called Registration Error Index (REI), which gives us a quantitative measure of registration error. The method achieves 0.882 AUC-ROC on the test dataset. Furthermore, the combined standard uncertainty of the estimated REI by our model lies within 11%, with a confidence level of approximately 68%.</p> <p>The second problem is about automatic radioisotope identification in real-time. A comparison of five different machine learning models is presented for gamma spectroscopy at a wide variety of testing conditions. Moreover, Hybrid Neural Network (HNN) was developed specifically for gamma-ray spectra data. Three different methods for feature extraction were tested as well. Experiments on MCNP simulated spectra suggest that HNN can achieve 2-12% higher F1 score at difficult testing conditions compared to best performing traditional ML models and obtains 93.33% F1 score during evaluation"--Abstract, page iii.</p>","abstract_has_math":false,"creators":["Galib, Shaikat Mahmood"],"institution":"Missouri University of Science and Technology","degree_name":"Ph. D. in Nuclear Engineering","degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":[],"committee_chairs":[],"committee_members":[],"year":null,"date_issued":"","date_published":null,"updated_at":"2026-07-24T03:18:34Z","subjects":["Deep learning","Machine learning","Medical image registration","Nuclear security","Radiation detection","Radiation imaging","Artificial Intelligence and Robotics","Bioimaging and Biomedical Optics","Nuclear"],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://scholarsmine.mst.edu/doctoral_dissertations/2829","outbound_label":"Repository record","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:creator","label":"Author","values":["Galib, Shaikat Mahmood"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:type","label":"Dc Type","values":["Dissertation - Open Access"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph. D. in Nuclear Engineering"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["Missouri University of Science and Technology"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Deep learning","Machine learning","Medical image registration","Nuclear security","Radiation detection","Radiation imaging","Artificial Intelligence and Robotics","Bioimaging and Biomedical Optics","Nuclear"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://scholarsmine.mst.edu/doctoral_dissertations/2829"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["\"The main focus of this work is to use machine learning and data mining techniques to address some challenging problems that arise from nuclear data. Specifically, two problem areas are discussed: nuclear imaging and radiation detection. The techniques to approach these problems are primarily based on a variant of Artificial Neural Network (ANN) called Convolutional Neural Network (CNN), which is one of the most popular forms of 'deep learning' technique.
The first problem is about interpreting and analyzing 3D medical radiation images automatically. A method is developed to identify and quantify deformable image registration (DIR) errors from lung CT scans for quality assurance (QA) purposes. The method includes the process of preparing a CT scan dataset for machine learning model training and testing, design of a 3D convolutional neural network architecture that classifies registrations into good or poor classes and a metric called Registration Error Index (REI), which gives us a quantitative measure of registration error. The method achieves 0.882 AUC-ROC on the test dataset. Furthermore, the combined standard uncertainty of the estimated REI by our model lies within 11%, with a confidence level of approximately 68%.
The second problem is about automatic radioisotope identification in real-time. A comparison of five different machine learning models is presented for gamma spectroscopy at a wide variety of testing conditions. Moreover, Hybrid Neural Network (HNN) was developed specifically for gamma-ray spectra data. Three different methods for feature extraction were tested as well. Experiments on MCNP simulated spectra suggest that HNN can achieve 2-12% higher F1 score at difficult testing conditions compared to best performing traditional ML models and obtains 93.33% F1 score during evaluation\"--Abstract, page iii.
"]},{"key":"dc:title","label":"Title","values":["Applications of machine learning in nuclear imaging and radiation detection"]}]}],"canonical_facts":{"dc:creator":["Galib, Shaikat Mahmood"],"dc:description.abstract":["\"The main focus of this work is to use machine learning and data mining techniques to address some challenging problems that arise from nuclear data. Specifically, two problem areas are discussed: nuclear imaging and radiation detection. The techniques to approach these problems are primarily based on a variant of Artificial Neural Network (ANN) called Convolutional Neural Network (CNN), which is one of the most popular forms of 'deep learning' technique.
The first problem is about interpreting and analyzing 3D medical radiation images automatically. A method is developed to identify and quantify deformable image registration (DIR) errors from lung CT scans for quality assurance (QA) purposes. The method includes the process of preparing a CT scan dataset for machine learning model training and testing, design of a 3D convolutional neural network architecture that classifies registrations into good or poor classes and a metric called Registration Error Index (REI), which gives us a quantitative measure of registration error. The method achieves 0.882 AUC-ROC on the test dataset. Furthermore, the combined standard uncertainty of the estimated REI by our model lies within 11%, with a confidence level of approximately 68%.
The second problem is about automatic radioisotope identification in real-time. A comparison of five different machine learning models is presented for gamma spectroscopy at a wide variety of testing conditions. Moreover, Hybrid Neural Network (HNN) was developed specifically for gamma-ray spectra data. Three different methods for feature extraction were tested as well. Experiments on MCNP simulated spectra suggest that HNN can achieve 2-12% higher F1 score at difficult testing conditions compared to best performing traditional ML models and obtains 93.33% F1 score during evaluation\"--Abstract, page iii.
"],"dc:identifier":["https://scholarsmine.mst.edu/doctoral_dissertations/2829"],"dc:subject":["Deep learning","Machine learning","Medical image registration","Nuclear security","Radiation detection","Radiation imaging","Artificial Intelligence and Robotics","Bioimaging and Biomedical Optics","Nuclear"],"dc:title":["Applications of machine learning in nuclear imaging and radiation detection"],"dc:type":["Dissertation - Open Access"],"thesis:degree_name":["Ph. D. in Nuclear Engineering"],"thesis:institution_name":["Missouri University of Science and Technology"]},"updated_at":"2026-07-24T03:18:34Z"}