NeAS: 3D modeling and surface extraction from X-ray images using Neural Attenuation Surface

• Item type: SIG Technical Reports(1)
• 公開日: 2024-01-18
• タイトル: NeAS: 3D modeling and surface extraction from X-ray images using Neural Attenuation Surface
• タイトル / 言語: en
• タイトル: NeAS: 3D modeling and surface extraction from X-ray images using Neural Attenuation Surface
• 言語: eng
• 資源タイプ / 資源タイプ識別子: http://purl.org/coar/resource_type/c_18gh
• 資源タイプ: technical report
• 著者所属: The University of Tokyo
• 著者所属: The University of Tokyo
• 著者所属: The University of Tokyo
• 著者所属: AIR WATER Inc.
• 著者所属: AIR WATER Inc.
• 著者所属: The University of Tokyo
• 著者所属(英): en
• 著者所属(英): The University of Tokyo
• 著者所属(英): en
• 著者所属(英): The University of Tokyo
• 著者所属(英): en
• 著者所属(英): The University of Tokyo
• 著者所属(英): en
• 著者所属(英): AIR WATER Inc.
• 著者所属(英): en
• 著者所属(英): AIR WATER Inc.
• 著者所属(英): en
• 著者所属(英): The University of Tokyo
• 著者名: Chengrui, Zhu; Ryoichi, Ishikawa; Masataka, Kagesawa; Tomohisa, Yuzawa; Toru, Watsuji; Takeshi, Oishi
• 著者名(英): Chengrui, Zhu; Ryoichi, Ishikawa; Masataka, Kagesawa; Tomohisa, Yuzawa; Toru, Watsuji; Takeshi, Oishi
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: Reconstructing 3D structures from 2D X-ray images is a valuable and efficient method in medical applications, offering the advantage of reducing patient radiation exposure compared to CT scans. We proposed NeAS, a novel approach for representing X-ray scene in a neural implicit way, an attenuation field is learnt to determine the spatial attenuation coefficients. To improve the 3D geometric accuracy of the results, NeAS also incorporates a learned signed distance function (SDF), which constrains the attenuation field and aids in extracting the 3D surface within the scene. Experiments were performed using both simulation and real X-ray images. The results demonstrate that when only 2D X-ray images are given, NeAS can accurately reconstruct 3D structures and extract surfaces within the scene accurately.
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: Reconstructing 3D structures from 2D X-ray images is a valuable and efficient method in medical applications, offering the advantage of reducing patient radiation exposure compared to CT scans. We proposed NeAS, a novel approach for representing X-ray scene in a neural implicit way, an attenuation field is learnt to determine the spatial attenuation coefficients. To improve the 3D geometric accuracy of the results, NeAS also incorporates a learned signed distance function (SDF), which constrains the attenuation field and aids in extracting the 3D surface within the scene. Experiments were performed using both simulation and real X-ray images. The results demonstrate that when only 2D X-ray images are given, NeAS can accurately reconstruct 3D structures and extract surfaces within the scene accurately.
• 書誌レコードID / 収録物識別子タイプ: NCID
• 書誌レコードID / 収録物識別子: AA11131797
• 書誌情報: 研究報告コンピュータビジョンとイメージメディア（CVIM）; 巻 2024-CVIM-236, 号 1, p. 1-9, 発行日 2024-01-18
• ISSN / 収録物識別子タイプ: ISSN
• ISSN / 収録物識別子: 2188-8701
• Notice: SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc.
• 出版者 / 言語: ja
• 出版者: 情報処理学会