{"id":{"repo_id":"marshall","oai_identifier":"oai:mds.marshall.edu:etd-2359"},"canonical_url":"https://search.ndltd.org/etd/marshall/oai:mds.marshall.edu:etd-2359","repository":{"repo_id":"marshall","name":"Marshall University","base_url":"https://mds.marshall.edu/do/oai/"},"display":{"title":"The U-Net-based Active Learning Framework for Enhancing Cancer Immunotherapy","abstract":"Breast cancer is the most common cancer in the world. According to the U.S. Breast Cancer Statistics, about 281,000 new cases of invasive breast cancer are expected to be diagnosed in 2021 (Smith et al., 2019). The death rate of breast cancer is higher than any other cancer type. Early detection and treatment of breast cancer have been challenging over the last few decades. Meanwhile, deep learning algorithms using Convolutional Neural Networks to segment images have achieved considerable success in recent years. These algorithms have continued to assist in exploring the quantitative measurement of cancer cells in the tumor microenvironment. However, detecting cancerous regions in whole-slide images has been challenging as it requires substantial annotation and training efforts from clinicians and biologists. In this thesis, a notable instructing process named U-Net-based Active Learning is proposed to improve the annotation and training procedure in a feedback learning process by utilizing a Deep Convolutional Neural Networks model. The proposed approach reduces the amount of time and effort required to analyze the whole slide images. During the Active Learning process, highly uncertain samples are iteratively selected to strategically supply characteristics of the whole slide images to the training process using a low-confidence sample selection algorithm. The performance results of the proposed approach indicated that the U-Net-based Active Learning framework has promising outcomes in the feedback learning process as it reaches 88.71% AUC-ROC when only using 64 image patches, while random lymphocyte prediction reaches 84.12% AUC-ROC at maximum.","abstract_html":"Breast cancer is the most common cancer in the world. According to the U.S. Breast Cancer Statistics, about 281,000 new cases of invasive breast cancer are expected to be diagnosed in 2021 (Smith et al., 2019). The death rate of breast cancer is higher than any other cancer type. Early detection and treatment of breast cancer have been challenging over the last few decades. Meanwhile, deep learning algorithms using Convolutional Neural Networks to segment images have achieved considerable success in recent years. These algorithms have continued to assist in exploring the quantitative measurement of cancer cells in the tumor microenvironment. However, detecting cancerous regions in whole-slide images has been challenging as it requires substantial annotation and training efforts from clinicians and biologists. In this thesis, a notable instructing process named U-Net-based Active Learning is proposed to improve the annotation and training procedure in a feedback learning process by utilizing a Deep Convolutional Neural Networks model. The proposed approach reduces the amount of time and effort required to analyze the whole slide images. During the Active Learning process, highly uncertain samples are iteratively selected to strategically supply characteristics of the whole slide images to the training process using a low-confidence sample selection algorithm. The performance results of the proposed approach indicated that the U-Net-based Active Learning framework has promising outcomes in the feedback learning process as it reaches 88.71% AUC-ROC when only using 64 image patches, while random lymphocyte prediction reaches 84.12% AUC-ROC at maximum.","abstract_has_math":false,"creators":["Joshi, Vishwanshi"],"institution":null,"degree_name":"Computer Science","degree_level":"Thesis","degree_discipline":"College of Engineering and Computer Sciences","degree_department":null,"school":null,"contributors":["Dr. Husnu Narman, Committee Chairperson","Dr. Sanghoon Lee","Dr. Haroon Malik"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":null,"date_issued":"","date_published":null,"updated_at":"2026-11-21T19:57:20Z","subjects":["Computer Engineering","Computer Sciences","Digital Communications and Networking","Diseases","Medicine and Health Sciences","Numerical Analysis and Scientific Computing","Oncology","OS and Networks","Physical Sciences and Mathematics","Active Learning","Deep Learning","Convolutional Neural Network","Image Processing","Whole-Slide Image"],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[],"person_identifier_entries":[]},"links":{"outbound_url":"https://mds.marshall.edu/etd/1352","outbound_label":"Repository record","outbound_source":"dc:identifier","record_url":"https://mds.marshall.edu/etd/1352","record_label":"Repository record","record_kind":"record","record_source":"dc:identifier","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://mds.marshall.edu/do/oai/?verb=GetRecord&metadataPrefix=oai_etdms&identifier=oai%3Amds.marshall.edu%3Aetd-2359","prefix":"oai_etdms"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Dr. Husnu Narman, Committee Chairperson","Dr. Sanghoon Lee","Dr. Haroon Malik"]},{"key":"dc:creator","label":"Author","values":["Joshi, Vishwanshi"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.available","label":"Dc Date Available","values":["2021-11-02T07:00:00Z"]},{"key":"dc:date.created","label":"Dc Date Created","values":["2021-01-01T08:00:00Z"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["College of Engineering and Computer Sciences"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Thesis"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Computer Science"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Computer Engineering","Computer Sciences","Digital Communications and Networking","Diseases","Medicine and Health Sciences","Numerical Analysis and Scientific Computing","Oncology","OS and Networks","Physical Sciences and Mathematics","Active Learning","Deep Learning","Convolutional Neural Network","Image Processing","Whole-Slide Image"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://mds.marshall.edu/etd/1352"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Breast cancer is the most common cancer in the world. According to the U.S. Breast Cancer Statistics, about 281,000 new cases of invasive breast cancer are expected to be diagnosed in 2021 (Smith et al., 2019). The death rate of breast cancer is higher than any other cancer type. Early detection and treatment of breast cancer have been challenging over the last few decades. Meanwhile, deep learning algorithms using Convolutional Neural Networks to segment images have achieved considerable success in recent years. These algorithms have continued to assist in exploring the quantitative measurement of cancer cells in the tumor microenvironment. However, detecting cancerous regions in whole-slide images has been challenging as it requires substantial annotation and training efforts from clinicians and biologists. In this thesis, a notable instructing process named U-Net-based Active Learning is proposed to improve the annotation and training procedure in a feedback learning process by utilizing a Deep Convolutional Neural Networks model. The proposed approach reduces the amount of time and effort required to analyze the whole slide images. During the Active Learning process, highly uncertain samples are iteratively selected to strategically supply characteristics of the whole slide images to the training process using a low-confidence sample selection algorithm. The performance results of the proposed approach indicated that the U-Net-based Active Learning framework has promising outcomes in the feedback learning process as it reaches 88.71% AUC-ROC when only using 64 image patches, while random lymphocyte prediction reaches 84.12% AUC-ROC at maximum."]},{"key":"dc:title","label":"Title","values":["The U-Net-based Active Learning Framework for Enhancing Cancer Immunotherapy"]}]}],"canonical_facts":{"dc:contributor":["Dr. Husnu Narman, Committee Chairperson","Dr. Sanghoon Lee","Dr. Haroon Malik"],"dc:creator":["Joshi, Vishwanshi"],"dc:date.available":["2021-11-02T07:00:00Z"],"dc:date.created":["2021-01-01T08:00:00Z"],"dc:description.abstract":["Breast cancer is the most common cancer in the world. According to the U.S. Breast Cancer Statistics, about 281,000 new cases of invasive breast cancer are expected to be diagnosed in 2021 (Smith et al., 2019). The death rate of breast cancer is higher than any other cancer type. Early detection and treatment of breast cancer have been challenging over the last few decades. Meanwhile, deep learning algorithms using Convolutional Neural Networks to segment images have achieved considerable success in recent years. These algorithms have continued to assist in exploring the quantitative measurement of cancer cells in the tumor microenvironment. However, detecting cancerous regions in whole-slide images has been challenging as it requires substantial annotation and training efforts from clinicians and biologists. In this thesis, a notable instructing process named U-Net-based Active Learning is proposed to improve the annotation and training procedure in a feedback learning process by utilizing a Deep Convolutional Neural Networks model. The proposed approach reduces the amount of time and effort required to analyze the whole slide images. During the Active Learning process, highly uncertain samples are iteratively selected to strategically supply characteristics of the whole slide images to the training process using a low-confidence sample selection algorithm. The performance results of the proposed approach indicated that the U-Net-based Active Learning framework has promising outcomes in the feedback learning process as it reaches 88.71% AUC-ROC when only using 64 image patches, while random lymphocyte prediction reaches 84.12% AUC-ROC at maximum."],"dc:identifier":["https://mds.marshall.edu/etd/1352"],"dc:subject":["Computer Engineering","Computer Sciences","Digital Communications and Networking","Diseases","Medicine and Health Sciences","Numerical Analysis and Scientific Computing","Oncology","OS and Networks","Physical Sciences and Mathematics","Active Learning","Deep Learning","Convolutional Neural Network","Image Processing","Whole-Slide Image"],"dc:title":["The U-Net-based Active Learning Framework for Enhancing Cancer Immunotherapy"],"thesis:degree_discipline":["College of Engineering and Computer Sciences"],"thesis:degree_level":["Thesis"],"thesis:degree_name":["Computer Science"]},"updated_at":"2026-11-21T19:57:20Z"}