{"id":{"repo_id":"calgary","oai_identifier":"oai:ucalgary.scholaris.ca:1880/118499"},"canonical_url":"https://search.dev.ndltd.org/etd/calgary/oai:ucalgary.scholaris.ca:1880/118499","repository":{"repo_id":"calgary","name":"University of Calgary","base_url":"https://ucalgary.scholaris.ca/server/oai/request"},"display":{"title":"Self-Supervised Learning Method for Semantic Segmentation of LiDAR Point Clouds","abstract":"Semantic segmentation has shown a significant success for achieving comprehensive scene understanding in real-time perception and urban modeling. Over the recent years, there have been significant advancements in semantic segmentation for LiDAR point clouds, largely the adopting of deep learning techniques. There are the related works of 3D semantic segmentation, including neural network models to process converted voxels, points, and graphs. However, point-based methods are not taken into account local structure feature, resulting in a lack of fine-grained features and limited generalization. Additionally, these models do not take full advantage of the high-level geometric correlations among local neighbors, resulting in low semantic segmentation accuracy. The use of voxel-based methods for balancing precision and computational efficiency is a useful technique. Still, voxel-based representation of point clouds is inefficient and tends to ignore fine details. Little research has investigated using graph-based methods for LiDAR point cloud semantic segmentation. Our work demonstrates the feasibility of using graph representation for highly accurate semantic segmentation in a point cloud. Other problems in the existing methods are high computational and memory requirements. Self-supervised learning on large unlabeled datasets is one way to reduce the number of manual annotations needed. In this thesis, we explore that leverage the combination self-supervised contrastive learning and graph-based method to overcome the semantic segmentation challenges of large-scale point clouds. An experimental qualitative and quantitative analysis of our method shows that the proposed approach can beat previous approaches on S3DIS and SemanticKITTI datasets for the task of LiDAR point cloud semantic segmentation.","abstract_html":"Semantic segmentation has shown a significant success for achieving comprehensive scene understanding in real-time perception and urban modeling. Over the recent years, there have been significant advancements in semantic segmentation for LiDAR point clouds, largely the adopting of deep learning techniques. There are the related works of 3D semantic segmentation, including neural network models to process converted voxels, points, and graphs. However, point-based methods are not taken into account local structure feature, resulting in a lack of fine-grained features and limited generalization. Additionally, these models do not take full advantage of the high-level geometric correlations among local neighbors, resulting in low semantic segmentation accuracy. The use of voxel-based methods for balancing precision and computational efficiency is a useful technique. Still, voxel-based representation of point clouds is inefficient and tends to ignore fine details. Little research has investigated using graph-based methods for LiDAR point cloud semantic segmentation. Our work demonstrates the feasibility of using graph representation for highly accurate semantic segmentation in a point cloud. Other problems in the existing methods are high computational and memory requirements. Self-supervised learning on large unlabeled datasets is one way to reduce the number of manual annotations needed. In this thesis, we explore that leverage the combination self-supervised contrastive learning and graph-based method to overcome the semantic segmentation challenges of large-scale point clouds. An experimental qualitative and quantitative analysis of our method shows that the proposed approach can beat previous approaches on S3DIS and SemanticKITTI datasets for the task of LiDAR point cloud semantic segmentation.","abstract_has_math":false,"creators":["Mutlu Kipirti, Fatma"],"institution":"Graduate Studies","degree_name":"Master of Science (MSc)","degree_level":null,"degree_discipline":"Engineering – Geomatics","degree_department":null,"school":null,"contributors":[],"advisors":["Wang, Ruisheng"],"committee_chairs":[],"committee_members":["Wang, Ruisheng","Hassan, Quazi Khalid","Wang, Xin","Yang, Hongzhou"],"year":2024,"date_issued":"2024-04-23","date_published":"2024-04-23","updated_at":"2026-07-24T01:30:36Z","subjects":["LiDAR point clouds","self-supervised learning","semantic segmentation"],"languages":["en"],"rights":["University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission."],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier.doi","label":"DOI","values":["https://doi.org/10.11575/PRISM/43341"],"render_values":[{"text":"https://doi.org/10.11575/PRISM/43341","href":"https://doi.org/10.11575/PRISM/43341","code":true}]}]},"links":{"outbound_url":"https://hdl.handle.net/1880/118499","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Wang, Ruisheng"]},{"key":"dc:contributor.committeemember","label":"Committee Member","values":["Wang, Ruisheng","Hassan, Quazi Khalid","Wang, Xin","Yang, Hongzhou"]},{"key":"dc:creator","label":"Author","values":["Mutlu Kipirti, Fatma"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2024-04-24T17:17:03Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2024-04-24T17:17:03Z"]},{"key":"dc:date.issued","label":"Date","values":["2024-04-23"]},{"key":"dc:publisher.institution","label":"Dc Publisher Institution","values":["University of Calgary"]},{"key":"dc:type","label":"Dc Type","values":["master thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Engineering – Geomatics"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Master of Science (MSc)"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Calgary"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["LiDAR point clouds","self-supervised learning","semantic segmentation"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission."]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.doi","label":"DOI","values":["https://doi.org/10.11575/PRISM/43341"]},{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/1880/118499"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Semantic segmentation has shown a significant success for achieving comprehensive scene understanding in real-time perception and urban modeling. Over the recent years, there have been significant advancements in semantic segmentation for LiDAR point clouds, largely the adopting of deep learning techniques. There are the related works of 3D semantic segmentation, including neural network models to process converted voxels, points, and graphs. However, point-based methods are not taken into account local structure feature, resulting in a lack of fine-grained features and limited generalization. Additionally, these models do not take full advantage of the high-level geometric correlations among local neighbors, resulting in low semantic segmentation accuracy. The use of voxel-based methods for balancing precision and computational efficiency is a useful technique. Still, voxel-based representation of point clouds is inefficient and tends to ignore fine details. Little research has investigated using graph-based methods for LiDAR point cloud semantic segmentation. Our work demonstrates the feasibility of using graph representation for highly accurate semantic segmentation in a point cloud. Other problems in the existing methods are high computational and memory requirements. Self-supervised learning on large unlabeled datasets is one way to reduce the number of manual annotations needed. In this thesis, we explore that leverage the combination self-supervised contrastive learning and graph-based method to overcome the semantic segmentation challenges of large-scale point clouds. An experimental qualitative and quantitative analysis of our method shows that the proposed approach can beat previous approaches on S3DIS and SemanticKITTI datasets for the task of LiDAR point cloud semantic segmentation."]},{"key":"dc:title","label":"Title","values":["Self-Supervised Learning Method for Semantic Segmentation of LiDAR Point Clouds"]}]}],"canonical_facts":{"dc:contributor.advisor":["Wang, Ruisheng"],"dc:contributor.committeemember":["Wang, Ruisheng","Hassan, Quazi Khalid","Wang, Xin","Yang, Hongzhou"],"dc:creator":["Mutlu Kipirti, Fatma"],"dc:date.accessioned":["2024-04-24T17:17:03Z"],"dc:date.available":["2024-04-24T17:17:03Z"],"dc:date.issued":["2024-04-23"],"dc:description.abstract":["Semantic segmentation has shown a significant success for achieving comprehensive scene understanding in real-time perception and urban modeling. Over the recent years, there have been significant advancements in semantic segmentation for LiDAR point clouds, largely the adopting of deep learning techniques. There are the related works of 3D semantic segmentation, including neural network models to process converted voxels, points, and graphs. However, point-based methods are not taken into account local structure feature, resulting in a lack of fine-grained features and limited generalization. Additionally, these models do not take full advantage of the high-level geometric correlations among local neighbors, resulting in low semantic segmentation accuracy. The use of voxel-based methods for balancing precision and computational efficiency is a useful technique. Still, voxel-based representation of point clouds is inefficient and tends to ignore fine details. Little research has investigated using graph-based methods for LiDAR point cloud semantic segmentation. Our work demonstrates the feasibility of using graph representation for highly accurate semantic segmentation in a point cloud. Other problems in the existing methods are high computational and memory requirements. Self-supervised learning on large unlabeled datasets is one way to reduce the number of manual annotations needed. In this thesis, we explore that leverage the combination self-supervised contrastive learning and graph-based method to overcome the semantic segmentation challenges of large-scale point clouds. An experimental qualitative and quantitative analysis of our method shows that the proposed approach can beat previous approaches on S3DIS and SemanticKITTI datasets for the task of LiDAR point cloud semantic segmentation."],"dc:identifier.doi":["https://doi.org/10.11575/PRISM/43341"],"dc:identifier.uri":["https://hdl.handle.net/1880/118499"],"dc:language.iso":["en"],"dc:publisher.institution":["University of Calgary"],"dc:rights":["University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission."],"dc:subject":["LiDAR point clouds","self-supervised learning","semantic segmentation"],"dc:title":["Self-Supervised Learning Method for Semantic Segmentation of LiDAR Point Clouds"],"dc:type":["master thesis"],"thesis:degree_discipline":["Engineering – Geomatics"],"thesis:degree_name":["Master of Science (MSc)"],"thesis:institution_name":["University of Calgary"]},"updated_at":"2026-07-24T01:30:36Z"}