True2Form: 3D Curve Networks from 2D Sketches via Selective Regularization

Baoxuan Xu; William Chang; Alla Sheffer; Adrien Bousseau; James McCrae; Karan Singh

首页> 外文期刊>ACM Transactions on Graphics >True2Form: 3D Curve Networks from 2D Sketches via Selective Regularization

【24h】

True2Form: 3D Curve Networks from 2D Sketches via Selective Regularization

机译：True2Form：通过选择性正则化从2D草图生成3D曲线网络

获取原文

获取原文并翻译 | 示例

掌桥外文数据库（机构版） >>

开具论文收录证明 >>

文献代查 >>

页面导航

摘要
著录项
相似文献
相关主题

摘要

True2Form is a sketch-based modeling system that reconstructs 3Drncurves from typical design sketches. Our approach to infer 3D formrnfrom 2D drawings is a novel mathematical framework of insightsrnderived from perception and design literature. We note that designersrnfavor viewpoints that maximally reveal 3D shape information,rnand strategically sketch descriptive curves that convey intrinsicrnshape properties, such as curvature, symmetry, or parallelism.rnStudies indicate that viewers apply these properties selectively tornenvision a globally consistent 3D shape. We mimic this selectivernregularization algorithmically, by progressively detecting and enforcingrnapplicable properties, accounting for their global impact onrnan evolving 3D curve network. Balancing regularity enforcementrnagainst sketch fidelity at each step allows us to correct for inaccuracyrninherent in free-hand sketching. We perceptually validaternour approach by showing agreement between our algorithm andrnviewers in selecting applicable regularities. We further evaluate ourrnsolution by: reconstructing a range of 3D models from diverselyrnsourced sketches; comparisons to prior art; and visual comparisonrnto both ground-truth and 3D reconstructions by designers.

机译：True2Form是一个基于草图的建模系统，可从典型设计草图中重建3Drncurves。我们从2D图纸推断3D形式的方法是一种新颖的数学框架，其洞察力来自于感知和设计文献。我们注意到，设计师偏爱的观点可以最大程度地揭示3D形状信息，并从战略上勾勒出传达固有形状属性（例如曲率，对称性或平行度）的描述性曲线。研究表明，观看者可以选择性地应用这些属性来设想全局一致的3D形状。我们通过逐步检测和强制应用属性，并考虑它们对3D曲线网络的全局影响，在算法上模拟这种选择性正则化。在每个步骤上使规则逼真度与草图逼真度保持平衡，这使我们能够纠正徒手草图绘制中固有的不准确性。通过在选择适用的规则上显示算法与查看者之间的一致性，我们在感知上验证了方法。我们通过以下方法来进一步评估解决方案：从不同来源的草图中重建一系列3D模型；与现有技术的比较；和视觉比较，由设计师进行地面和3D重建。

著录项

来源
《ACM Transactions on Graphics》 |2014年第4cd期|131.1-131.13|共13页
作者
Baoxuan Xu; William Chang; Alla Sheffer; Adrien Bousseau; James McCrae; Karan Singh;
展开▼
作者单位

University of British Columbia;

University of British Columbia;

University of British Columbia;

Inria;

University of Toronto;

University of Toronto;

展开▼
收录信息美国《科学引文索引》(SCI);
原文格式 PDF
正文语种 eng
中图分类
关键词
sketch-based modeling; descriptive lines; regularity;

机译：基于草图的建模;描述性线条;规律性;

相似文献

外文文献
中文文献
专利

1. Building 3D surface networks from 2D curve networks with application to anatomical modeling [J] . Tao Ju, Joe Warren, James Carson, The Visual Computer . 2005,第8a10期

机译：从2D曲线网络构建3D曲面网络并将其应用于解剖建模
2. Modeling piecewise helix curves from 2D sketches [J] . Nicolas Cherin, Frederic Cordier, Mahmoud Melkemi Computer-Aided Design . 2014,第Null期

机译：从2D草图建模分段螺旋曲线
3. Automated Hand-drawn sketches retrieval and recognition using regularized Particle Swarm Optimization based deep convolutional neural network [J] . Thangakrishnan M. Suresh, Ramar K. Journal of ambient intelligence and humanized computing . 2021,第6期

机译：自动化手绘草图检索和识别使用正则化粒子群优化基于深度卷积神经网络
4. Automatic 3D Modelling for Heel Based on the Images of 2D Sketches [C] . Chao Wu, Jiping Li, Wanrong Gu, International Conference on Advanced Electronic Materials, Computers and Software Engineering . 2020

机译：基于2D草图图像的鞋跟自动3D建模
5. 2D and 3D Computational Optical Imaging Using Deep Convolutional Neural Networks (DCNNs) [D] . Nguyen, Thanh 2019

机译：使用深度卷积神经网络（DCNN）的2D和3D计算光学成像
6. 3D Convolutional Neural Networks Initialized from Pretrained 2D Convolutional Neural Networks for Classification of Industrial Parts [O] . Ibon Merino, Jon Azpiazu, Anthony Remazeilles, 2021

机译：3D卷积神经网络从佩带的2D卷积神经网络初始化用于工业部件的分类
7. Modelling 3D Objects Using 2D Sketches Through Radial Renderings of Curvature Maps [O] . Frode Eika Sandnes, Evelyn Eika 2018

机译：使用2D草图建模3D对象通过曲率图的径向渲染

True2Form: 3D Curve Networks from 2D Sketches via Selective Regularization

摘要

著录项

相似文献

相关主题

期刊订阅