Attention-Guided Multi-scale Neural Dual Contouring

Simple item page
dc.contributor.authorWu, Fulien_US
dc.contributor.authorHu, Chaoranen_US
dc.contributor.authorLi, Wenxuanen_US
dc.contributor.authorHao, Pengyien_US
dc.contributor.editorChristie, Marcen_US
dc.contributor.editorHan, Ping-Hsuanen_US
dc.contributor.editorLin, Shih-Syunen_US
dc.contributor.editorPietroni, Nicoen_US
dc.contributor.editorSchneider, Teseoen_US
dc.contributor.editorTsai, Hsin-Rueyen_US
dc.contributor.editorWang, Yu-Shuenen_US
dc.contributor.editorZhang, Eugeneen_US
dc.date.accessioned2025-10-07T06:04:21Z
dc.date.available2025-10-07T06:04:21Z
dc.date.issued2025
dc.description.abstractReconstructing high-quality meshes from binary voxel data is a fundamental task in computer graphics. However, existing methods struggle with low information density and strong discreteness, making it difficult to capture complex geometry and long-range boundary features, often leading to jagged surfaces and loss of sharp details.We propose an Attention-Guided Multiscale Neural Dual Contouring (AGNDC) method to address this challenge. AGNDC refines surface reconstruction through a multi-scale framework, using a hybrid feature extractor that combines global attention and dynamic snake convolution to enhance perception of long-range and high-curvature features. A dynamic feature fusion module aligns multi-scale predictions to improve local detail continuity, while a geometric postprocessing module further refines mesh boundaries and suppresses artifacts. Experiments on the ABC dataset demonstrate the superior performance of AGNDC in both visual and quantitative metrics. It achieves a Chamfer Distance (CD×105) of 9.013 and an F-score of 0.440, significantly reducing jaggedness and improving surface smoothness.en_US
dc.description.sectionheaders3D Reconstruction
dc.description.seriesinformationPacific Graphics Conference Papers, Posters, and Demos
dc.identifier.doi10.2312/pg.20251291
dc.identifier.isbn978-3-03868-295-0
dc.identifier.pages12 pages
dc.identifier.urihttps://doi.org/10.2312/pg.20251291
dc.identifier.urihttps://diglib.eg.org/handle/10.2312/pg20251291
dc.publisherThe Eurographics Associationen_US
dc.rightsAttribution 4.0 International License
dc.rights.urihttps://creativecommons.org/licenses/by/4.0/
dc.subjectCCS Concepts: Computing methodologies → Mesh generation; Neural networks; Volumetric models
dc.subjectComputing methodologies → Mesh generation
dc.subjectNeural networks
dc.subjectVolumetric models
dc.titleAttention-Guided Multi-scale Neural Dual Contouringen_US
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
pg20251291.pdf
Size:
9.12 MB
Format:
Adobe Portable Document Format
Download
Collections
PG2025 Conference Papers, Posters, and Demos