{"id":{"repo_id":"adelaide","oai_identifier":"oai:digital.library.adelaide.edu.au:2440/136692"},"canonical_url":"https://search.dev.ndltd.org/etd/adelaide/oai:digital.library.adelaide.edu.au:2440/136692","repository":{"repo_id":"adelaide","name":"University of Adelaide","base_url":"https://digital.library.adelaide.edu.au/server/oai/request"},"display":{"title":"Self-supervised Learning of Monocular Depth from Video","abstract":"Image-based depth estimation as a fundamental problem in computer vision allows for understanding the scene geometry using only cameras. This thesis addresses the specific problem of monocular depth estimation via self-supervised learning from RGB-only videos. Although existing work has shown partial excellent results in benchmark datasets, there remain several vital challenges that limit the use of these algorithms in general scenarios. To summarize, my identified challenges and contributions include: (i) Previous methods predict inconsistent depths over a video, which limits their uses in visual localization and mapping. To this end, I propose a geometry consistency loss that penalizes the multi-view depth misalignment in training, which enables scale-consistent depth estimation at inference time; (ii) Previous methods often diverge or show low-accuracy results when training on handheld camera captured videos. To address the challenge, I analyze the effect of camera motion on depth network gradients, and I propose an auto-rectify network to remove the relative rotation in training image pairs for robust learning; (iii) Previous methods fail to learn reasonable depths from highly dynamic scenes due to the non-rigidity. In this scenario, I propose a novel method, which constrains dynamic regions using an external well-trained depth estimation network and supervises static regions via multi-view losses. Comprehensive quantitative results and rich qualitative results are provided to demonstrate the advantages of the proposed methods over existing alternatives. The codes and pre-trained models have been released at https://github.com/JiawangBian","abstract_html":"Image-based depth estimation as a fundamental problem in computer vision allows for understanding the scene geometry using only cameras. This thesis addresses the specific problem of monocular depth estimation via self-supervised learning from RGB-only videos. Although existing work has shown partial excellent results in benchmark datasets, there remain several vital challenges that limit the use of these algorithms in general scenarios. To summarize, my identified challenges and contributions include: (i) Previous methods predict inconsistent depths over a video, which limits their uses in visual localization and mapping. To this end, I propose a geometry consistency loss that penalizes the multi-view depth misalignment in training, which enables scale-consistent depth estimation at inference time; (ii) Previous methods often diverge or show low-accuracy results when training on handheld camera captured videos. To address the challenge, I analyze the effect of camera motion on depth network gradients, and I propose an auto-rectify network to remove the relative rotation in training image pairs for robust learning; (iii) Previous methods fail to learn reasonable depths from highly dynamic scenes due to the non-rigidity. In this scenario, I propose a novel method, which constrains dynamic regions using an external well-trained depth estimation network and supervises static regions via multi-view losses. Comprehensive quantitative results and rich qualitative results are provided to demonstrate the advantages of the proposed methods over existing alternatives. The codes and pre-trained models have been released at https://github.com/JiawangBian","abstract_has_math":false,"creators":["Bian, Jiawang"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["Reid, Ian","Shen, Chunhua (Zhejiang University)"],"committee_chairs":[],"committee_members":[],"year":2022,"date_issued":"2022","date_published":"2022","updated_at":"2026-07-24T00:50:39Z","subjects":["Monocular Depth Estimation, Self-supervised Learning, Visual SLAM"],"languages":["en"],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/2440/136692","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Reid, Ian","Shen, Chunhua (Zhejiang University)"]},{"key":"dc:creator","label":"Author","values":["Bian, Jiawang"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.issued","label":"Date","values":["2022"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Monocular Depth Estimation, Self-supervised Learning, Visual SLAM"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/2440/136692"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Image-based depth estimation as a fundamental problem in computer vision allows for understanding the scene geometry using only cameras. This thesis addresses the specific problem of monocular depth estimation via self-supervised learning from RGB-only videos. Although existing work has shown partial excellent results in benchmark datasets, there remain several vital challenges that limit the use of these algorithms in general scenarios. To summarize, my identified challenges and contributions include: (i) Previous methods predict inconsistent depths over a video, which limits their uses in visual localization and mapping. To this end, I propose a geometry consistency loss that penalizes the multi-view depth misalignment in training, which enables scale-consistent depth estimation at inference time; (ii) Previous methods often diverge or show low-accuracy results when training on handheld camera captured videos. To address the challenge, I analyze the effect of camera motion on depth network gradients, and I propose an auto-rectify network to remove the relative rotation in training image pairs for robust learning; (iii) Previous methods fail to learn reasonable depths from highly dynamic scenes due to the non-rigidity. In this scenario, I propose a novel method, which constrains dynamic regions using an external well-trained depth estimation network and supervises static regions via multi-view losses. Comprehensive quantitative results and rich qualitative results are provided to demonstrate the advantages of the proposed methods over existing alternatives. The codes and pre-trained models have been released at https://github.com/JiawangBian"]},{"key":"dc:title","label":"Title","values":["Self-supervised Learning of Monocular Depth from Video"]}]}],"canonical_facts":{"dc:contributor.advisor":["Reid, Ian","Shen, Chunhua (Zhejiang University)"],"dc:creator":["Bian, Jiawang"],"dc:date.issued":["2022"],"dc:description.abstract":["Image-based depth estimation as a fundamental problem in computer vision allows for understanding the scene geometry using only cameras. This thesis addresses the specific problem of monocular depth estimation via self-supervised learning from RGB-only videos. Although existing work has shown partial excellent results in benchmark datasets, there remain several vital challenges that limit the use of these algorithms in general scenarios. To summarize, my identified challenges and contributions include: (i) Previous methods predict inconsistent depths over a video, which limits their uses in visual localization and mapping. To this end, I propose a geometry consistency loss that penalizes the multi-view depth misalignment in training, which enables scale-consistent depth estimation at inference time; (ii) Previous methods often diverge or show low-accuracy results when training on handheld camera captured videos. To address the challenge, I analyze the effect of camera motion on depth network gradients, and I propose an auto-rectify network to remove the relative rotation in training image pairs for robust learning; (iii) Previous methods fail to learn reasonable depths from highly dynamic scenes due to the non-rigidity. In this scenario, I propose a novel method, which constrains dynamic regions using an external well-trained depth estimation network and supervises static regions via multi-view losses. Comprehensive quantitative results and rich qualitative results are provided to demonstrate the advantages of the proposed methods over existing alternatives. The codes and pre-trained models have been released at https://github.com/JiawangBian"],"dc:identifier.uri":["https://hdl.handle.net/2440/136692"],"dc:language.iso":["en"],"dc:subject":["Monocular Depth Estimation, Self-supervised Learning, Visual SLAM"],"dc:title":["Self-supervised Learning of Monocular Depth from Video"],"dc:type":["Thesis"]},"updated_at":"2026-07-24T00:50:39Z"}