Saved in:
Bibliographic Details
Main Authors: Yao, Yuxin, Ren, Siyu, Hou, Junhui, Deng, Zhi, Zhang, Juyong, Wang, Wenping
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2403.11586
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909264264560640
author Yao, Yuxin
Ren, Siyu
Hou, Junhui
Deng, Zhi
Zhang, Juyong
Wang, Wenping
author_facet Yao, Yuxin
Ren, Siyu
Hou, Junhui
Deng, Zhi
Zhang, Juyong
Wang, Wenping
contents This paper explores the problem of reconstructing temporally consistent surfaces from a 3D point cloud sequence without correspondence. To address this challenging task, we propose DynoSurf, an unsupervised learning framework integrating a template surface representation with a learnable deformation field. Specifically, we design a coarse-to-fine strategy for learning the template surface based on the deformable tetrahedron representation. Furthermore, we propose a learnable deformation representation based on the learnable control points and blending weights, which can deform the template surface non-rigidly while maintaining the consistency of the local shape. Experimental results demonstrate the significant superiority of DynoSurf over current state-of-the-art approaches, showcasing its potential as a powerful tool for dynamic mesh reconstruction. The code is publicly available at https://github.com/yaoyx689/DynoSurf.
format Preprint
id arxiv_https___arxiv_org_abs_2403_11586
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle DynoSurf: Neural Deformation-based Temporally Consistent Dynamic Surface Reconstruction
Yao, Yuxin
Ren, Siyu
Hou, Junhui
Deng, Zhi
Zhang, Juyong
Wang, Wenping
Computer Vision and Pattern Recognition
This paper explores the problem of reconstructing temporally consistent surfaces from a 3D point cloud sequence without correspondence. To address this challenging task, we propose DynoSurf, an unsupervised learning framework integrating a template surface representation with a learnable deformation field. Specifically, we design a coarse-to-fine strategy for learning the template surface based on the deformable tetrahedron representation. Furthermore, we propose a learnable deformation representation based on the learnable control points and blending weights, which can deform the template surface non-rigidly while maintaining the consistency of the local shape. Experimental results demonstrate the significant superiority of DynoSurf over current state-of-the-art approaches, showcasing its potential as a powerful tool for dynamic mesh reconstruction. The code is publicly available at https://github.com/yaoyx689/DynoSurf.
title DynoSurf: Neural Deformation-based Temporally Consistent Dynamic Surface Reconstruction
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2403.11586