3D shape analysis via multi-modal contrastive learning

IF 1.3 4区 计算机科学 Q3 COMPUTER SCIENCE, SOFTWARE ENGINEERING
Zhenyu Shu , Xufei Sun , Chaoyi Pang
{"title":"3D shape analysis via multi-modal contrastive learning","authors":"Zhenyu Shu ,&nbsp;Xufei Sun ,&nbsp;Chaoyi Pang","doi":"10.1016/j.cagd.2025.102454","DOIUrl":null,"url":null,"abstract":"<div><div>In recent years, 3D shape analysis has emerged as a crucial field with applications in various domains, such as multimedia processing, computer graphics, computer vision, and robotics. The ability to understand and interpret 3D shapes is fundamental for tasks like 3D shape segmentation, points of interest detection, shape retrieval, recognition, and generation. However, the complexity of 3D mesh models is a significant barrier that stops the topic from enhancing. Thus, we propose a novel 3D shape analysis framework in this paper by multi-modal contrastive learning techniques. Our framework makes use of the original mesh data and the projected images from various points of view of the mesh model. Those two modals contribute to providing more precise features with the help of our within-modal and cross-modal losses, which respectively calculate the distances of feature vectors within the mesh model and between feature vectors of mesh and image. Our framework is tested on downstream tasks, including 3D shape segmentation and points of interest detection, and outperforms most state-of-the-art methods on public datasets.</div></div>","PeriodicalId":55226,"journal":{"name":"Computer Aided Geometric Design","volume":"119 ","pages":"Article 102454"},"PeriodicalIF":1.3000,"publicationDate":"2025-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Computer Aided Geometric Design","FirstCategoryId":"94","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0167839625000433","RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"COMPUTER SCIENCE, SOFTWARE ENGINEERING","Score":null,"Total":0}
引用次数: 0

Abstract

In recent years, 3D shape analysis has emerged as a crucial field with applications in various domains, such as multimedia processing, computer graphics, computer vision, and robotics. The ability to understand and interpret 3D shapes is fundamental for tasks like 3D shape segmentation, points of interest detection, shape retrieval, recognition, and generation. However, the complexity of 3D mesh models is a significant barrier that stops the topic from enhancing. Thus, we propose a novel 3D shape analysis framework in this paper by multi-modal contrastive learning techniques. Our framework makes use of the original mesh data and the projected images from various points of view of the mesh model. Those two modals contribute to providing more precise features with the help of our within-modal and cross-modal losses, which respectively calculate the distances of feature vectors within the mesh model and between feature vectors of mesh and image. Our framework is tested on downstream tasks, including 3D shape segmentation and points of interest detection, and outperforms most state-of-the-art methods on public datasets.

Abstract Image

基于多模态对比学习的三维形状分析
近年来,三维形状分析已成为多媒体处理、计算机图形学、计算机视觉和机器人技术等各个领域的重要应用领域。理解和解释3D形状的能力是诸如3D形状分割、兴趣点检测、形状检索、识别和生成等任务的基础。然而,三维网格模型的复杂性是阻碍该主题增强的重要障碍。因此,本文提出了一种基于多模态对比学习技术的三维形状分析框架。我们的框架利用了原始网格数据和从网格模型的各个角度投影的图像。这两种模态有助于提供更精确的特征,我们的模态内损失和跨模态损失分别计算网格模型内特征向量的距离和网格与图像特征向量之间的距离。我们的框架在下游任务上进行了测试,包括3D形状分割和兴趣点检测,并且在公共数据集上优于大多数最先进的方法。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
Computer Aided Geometric Design
Computer Aided Geometric Design 工程技术-计算机:软件工程
CiteScore
3.50
自引率
13.30%
发文量
57
审稿时长
60 days
期刊介绍: The journal Computer Aided Geometric Design is for researchers, scholars, and software developers dealing with mathematical and computational methods for the description of geometric objects as they arise in areas ranging from CAD/CAM to robotics and scientific visualization. The journal publishes original research papers, survey papers and with quick editorial decisions short communications of at most 3 pages. The primary objects of interest are curves, surfaces, and volumes such as splines (NURBS), meshes, subdivision surfaces as well as algorithms to generate, analyze, and manipulate them. This journal will report on new developments in CAGD and its applications, including but not restricted to the following: -Mathematical and Geometric Foundations- Curve, Surface, and Volume generation- CAGD applications in Numerical Analysis, Computational Geometry, Computer Graphics, or Computer Vision- Industrial, medical, and scientific applications. The aim is to collect and disseminate information on computer aided design in one journal. To provide the user community with methods and algorithms for representing curves and surfaces. To illustrate computer aided geometric design by means of interesting applications. To combine curve and surface methods with computer graphics. To explain scientific phenomena by means of computer graphics. To concentrate on the interaction between theory and application. To expose unsolved problems of the practice. To develop new methods in computer aided geometry.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信