Computer-Aided Design最新文献_第9页

Toward precise curve offsetting constrained to parametric surfaces 向着受参数曲面约束的精确曲线偏移

IF 3.1 3区计算机科学 Q2 COMPUTER SCIENCE, SOFTWARE ENGINEERING

Computer-Aided Design

Pub Date : 2025-07-21 DOI: 10.1016/j.cad.2025.103916

Jin Zhao , Pengfei Wang , Shuangmin Chen , Jiong Guo , Shiqing Xin , Changhe Tu , Wenping Wang

Computing offsets of curves on parametric surfaces is a fundamental yet challenging operation in computer-aided design and manufacturing. Traditional analytical approaches suffer from time-consuming geodesic distance queries and complex self-intersection handling, while discrete methods often struggle with precision. In this paper, we propose a totally different algorithm paradigm. Our key insight is that by representing the source curve as a sequence of line-segment primitives, the Voronoi decomposition constrained to the parametric surface enables localized offset computation. Specifically, the offsetting process can be efficiently traced by independently visiting the corresponding Voronoi cells. To address the challenge of computing the Voronoi decomposition on parametric surfaces, we introduce two key techniques. First, we employ intrinsic triangulation in the parameter space to accurately capture geodesic distances. Second, instead of directly computing the surface-constrained Voronoi decomposition, we decompose the triangulated parameter plane using a series of plane-cutting operations. Experimental results demonstrate that our algorithm achieves superior accuracy and runtime performance compared to existing methods. We also present several practical applications enabled by our approach.

在计算机辅助设计和制造中，计算参数化曲面上曲线的偏移量是一项基本但具有挑战性的操作。传统的分析方法受到耗时的测地线距离查询和复杂的自交处理的困扰，而离散方法往往难以达到精度。在本文中，我们提出了一个完全不同的算法范式。我们的关键见解是，通过将源曲线表示为线段原语序列，Voronoi分解约束于参数表面，可以实现局部偏移计算。具体来说，通过独立访问相应的Voronoi细胞，可以有效地跟踪抵消过程。为了解决在参数曲面上计算Voronoi分解的挑战，我们引入了两个关键技术。首先，我们在参数空间中使用内禀三角剖分来精确捕获测地线距离。其次，我们使用一系列的平面切割操作来分解三角化参数平面，而不是直接计算表面约束的Voronoi分解。实验结果表明，与现有方法相比，该算法具有更高的精度和运行时间性能。我们还介绍了通过我们的方法实现的几个实际应用。

{"title":"Toward precise curve offsetting constrained to parametric surfaces","authors":"Jin Zhao , Pengfei Wang , Shuangmin Chen , Jiong Guo , Shiqing Xin , Changhe Tu , Wenping Wang","doi":"10.1016/j.cad.2025.103916","DOIUrl":"10.1016/j.cad.2025.103916","url":null,"abstract":"<div><div>Computing offsets of curves on parametric surfaces is a fundamental yet challenging operation in computer-aided design and manufacturing. Traditional analytical approaches suffer from time-consuming geodesic distance queries and complex self-intersection handling, while discrete methods often struggle with precision. In this paper, we propose a totally different algorithm paradigm. Our key insight is that by representing the source curve as a sequence of line-segment primitives, the Voronoi decomposition constrained to the parametric surface enables localized offset computation. Specifically, the offsetting process can be efficiently traced by independently visiting the corresponding Voronoi cells. To address the challenge of computing the Voronoi decomposition on parametric surfaces, we introduce two key techniques. First, we employ intrinsic triangulation in the parameter space to accurately capture geodesic distances. Second, instead of directly computing the surface-constrained Voronoi decomposition, we decompose the triangulated parameter plane using a series of plane-cutting operations. Experimental results demonstrate that our algorithm achieves superior accuracy and runtime performance compared to existing methods. We also present several practical applications enabled by our approach.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"188 ","pages":"Article 103916"},"PeriodicalIF":3.1,"publicationDate":"2025-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144721019","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

UCSM: Dataset of U-shaped parametric CAD geometries and real-world sheet metal meshes for deep drawing UCSM: u形参数化CAD几何数据集和用于深拉深的真实金属板网格

IF 3 3区计算机科学 Q2 COMPUTER SCIENCE, SOFTWARE ENGINEERING

Computer-Aided Design

Pub Date : 2025-07-18 DOI: 10.1016/j.cad.2025.103924

Tobias Lehrer , Philipp Stocker , Fabian Duddeck , Marcus Wagner

The development of machine learning (ML) applications in deep drawing is hindered by limited data availability and the absence of open-access benchmarks for validating novel approaches, including domain generalization over distinct geometries. This paper addresses these challenges by introducing a comprehensive U-shaped dataset tailored to this manufacturing process. Our U-Channel sheet metal (UCSM) dataset combines 90 real-world meshes with an infinite number of synthetic geometry samples generated from four parametric Computer-Aided Design (CAD) models, ensuring extensive geometry variety and data quantity. Additionally, a ready-to-use dataset for drawability assessment and segmentation is provided. Leveraging CAD and mesh data sources bridges the gap between sparse data availability and ML requirements. Our analysis demonstrates that the proposed parametric models are geometrically valid, and real-world and synthetic data complement each other effectively, providing robust support for ML model development. While the dataset is confined to U-shaped, thin-walled, deep drawing scenarios, it considerably aids in overcoming data scarcity. Thereby, it facilitates the validation and comparison of new geometry-generalizing ML methodologies in this domain. By providing this benchmark dataset, we enhance the comparability and validation of emerging methods for ML advancements in sheet metal forming.

由于有限的数据可用性和缺乏用于验证新方法的开放访问基准，包括不同几何形状的域泛化，机器学习（ML）在深拉深中的应用的发展受到阻碍。本文通过引入针对该制造过程量身定制的全面u形数据集来解决这些挑战。我们的U-Channel金属板（UCSM）数据集结合了90个真实世界的网格和由四个参数化计算机辅助设计（CAD）模型生成的无限数量的合成几何样本，确保了广泛的几何多样性和数据量。此外，还提供了可绘制性评估和分割的现成数据集。利用CAD和网格数据源弥合了稀疏数据可用性和ML需求之间的差距。我们的分析表明，所提出的参数模型在几何上是有效的，并且真实世界和合成数据有效地相互补充，为ML模型的开发提供了强大的支持。虽然数据集仅限于u形、薄壁、深拉伸场景，但它在很大程度上有助于克服数据稀缺性。因此，它促进了该领域中新的几何泛化ML方法的验证和比较。通过提供这个基准数据集，我们增强了在钣金成形中ML进步的新兴方法的可比性和有效性。

{"title":"UCSM: Dataset of U-shaped parametric CAD geometries and real-world sheet metal meshes for deep drawing","authors":"Tobias Lehrer , Philipp Stocker , Fabian Duddeck , Marcus Wagner","doi":"10.1016/j.cad.2025.103924","DOIUrl":"10.1016/j.cad.2025.103924","url":null,"abstract":"<div><div>The development of machine learning (ML) applications in deep drawing is hindered by limited data availability and the absence of open-access benchmarks for validating novel approaches, including domain generalization over distinct geometries. This paper addresses these challenges by introducing a comprehensive U-shaped dataset tailored to this manufacturing process. Our U-Channel sheet metal (UCSM) dataset combines 90 real-world meshes with an infinite number of synthetic geometry samples generated from four parametric Computer-Aided Design (CAD) models, ensuring extensive geometry variety and data quantity. Additionally, a ready-to-use dataset for drawability assessment and segmentation is provided. Leveraging CAD and mesh data sources bridges the gap between sparse data availability and ML requirements. Our analysis demonstrates that the proposed parametric models are geometrically valid, and real-world and synthetic data complement each other effectively, providing robust support for ML model development. While the dataset is confined to U-shaped, thin-walled, deep drawing scenarios, it considerably aids in overcoming data scarcity. Thereby, it facilitates the validation and comparison of new geometry-generalizing ML methodologies in this domain. By providing this benchmark dataset, we enhance the comparability and validation of emerging methods for ML advancements in sheet metal forming.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"188 ","pages":"Article 103924"},"PeriodicalIF":3.0,"publicationDate":"2025-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144713180","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Computational design and fabrication of reusable multi-tangent bar structures 可重复使用多切线杆结构的计算设计与制造

IF 3 3区计算机科学 Q2 COMPUTER SCIENCE, SOFTWARE ENGINEERING

Computer-Aided Design

Pub Date : 2025-07-17 DOI: 10.1016/j.cad.2025.103907

Yijiang Huang , Ziqi Wang , Yi-Hsiu Hung , Chenming Jiang , Aurèle L. Gheyselinck , Stelian Coros

Temporary bar structures made of reusable standardized components are widely used in construction, events, and exhibitions. They are economical, easy to assemble, and can be disassembled and reused in various structural arrangements for various purposes. However, existing reusable temporary structures are either limited to modular yet repetitive designs or require bespoke components, which restricts their reuse potential. Instead of designing bespoke kit of parts for limited reuse, this paper investigates how to design and build diverse freeform structures from one homogeneous kit of parts. We propose a computational framework to generate multi-tangent bar structures, a widely used jointing system, which allows bars to be joined at any point along their length with standard connectors. We present a mathematical formulation and a numerical scheme to optimize the bar spatial positions and contact assignment simultaneously, while ensuring that the constraints of tangency, collision, joint connectivity, and bar length are satisfied. Together with simulated case studies, we present two physical prototypes that reuse the same kit of parts using an augmented reality-guided assembly workflow.

由可重复使用的标准化构件制成的临时酒吧结构广泛应用于建筑、活动和展览中。它们经济，易于组装，并且可以拆卸和重复使用各种结构安排，用于各种用途。然而，现有的可重用临时结构要么局限于模块化的重复设计，要么需要定制的组件，这限制了它们的重用潜力。本文研究了如何从一个均匀的零件套件设计和构建不同的自由形状结构，而不是为有限的重用而设计定制的零件套件。我们提出了一个计算框架来生成多切线杆结构，这是一种广泛使用的连接系统，它允许沿其长度的任何点使用标准连接器连接杆。在保证切线、碰撞、关节连通性和杆长约束的前提下，提出了优化杆空间位置和接触分配的数学公式和数值方案。与模拟案例研究一起，我们提出了两个物理原型，使用增强现实引导的装配工作流程重用相同的部件套件。

{"title":"Computational design and fabrication of reusable multi-tangent bar structures","authors":"Yijiang Huang , Ziqi Wang , Yi-Hsiu Hung , Chenming Jiang , Aurèle L. Gheyselinck , Stelian Coros","doi":"10.1016/j.cad.2025.103907","DOIUrl":"10.1016/j.cad.2025.103907","url":null,"abstract":"<div><div>Temporary bar structures made of reusable standardized components are widely used in construction, events, and exhibitions. They are economical, easy to assemble, and can be disassembled and reused in various structural arrangements for various purposes. However, existing reusable temporary structures are either limited to modular yet repetitive designs or require bespoke components, which restricts their reuse potential. Instead of designing bespoke kit of parts for limited reuse, this paper investigates how to design and build diverse freeform structures from one homogeneous kit of parts. We propose a computational framework to generate multi-tangent bar structures, a widely used jointing system, which allows bars to be joined at any point along their length with standard connectors. We present a mathematical formulation and a numerical scheme to optimize the bar spatial positions and contact assignment simultaneously, while ensuring that the constraints of tangency, collision, joint connectivity, and bar length are satisfied. Together with simulated case studies, we present two physical prototypes that reuse the same kit of parts using an augmented reality-guided assembly workflow.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"188 ","pages":"Article 103907"},"PeriodicalIF":3.0,"publicationDate":"2025-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144702899","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Implicit complex topology modeling design method based on Voronoi 基于Voronoi的隐式复杂拓扑建模设计方法

IF 3 3区计算机科学 Q2 COMPUTER SCIENCE, SOFTWARE ENGINEERING

Computer-Aided Design

Pub Date : 2025-07-17 DOI: 10.1016/j.cad.2025.103910

Chongshui Liu, Quan Qi

Porous structures with complex topologies have long been at the forefront of research in the field of additive manufacturing. However, existing methods exhibit limited capability in representing the morphology of pores in random porous structures, lack precision in gradient design, and struggle to accurately simulate the stochastic porous features found in nature. This paper proposes a novel modeling approach for random porous structures. Firstly, the method utilizes Voronoi tessellation to obtain the topological data of three-dimensional Voronoi diagrams. Based on this topology, a signed distance field is constructed to implicitly represent the porous model, providing exceptional flexibility in the expression of pore morphologies. Secondly, growth functions that describe natural growth processes are employed to achieve smooth transitions in pore morphology within the porous model, enabling precise control over transitions and facilitating the design of hierarchical and gradient porous structures. Finally, by introducing a radial basis function (RBF) weighted interpolation method, the transition region is generalized to the entire domain, allowing for complex gradient designs and biomimetic simulations. This study offers a new solution for the modeling of complex topologies.

具有复杂拓扑结构的多孔结构一直是增材制造领域研究的前沿。然而，现有的方法在表征随机多孔结构中孔隙形态的能力有限，在梯度设计方面缺乏精度，并且难以准确模拟自然界中发现的随机多孔特征。本文提出了一种新的随机多孔结构的建模方法。该方法首先利用Voronoi镶嵌获得三维Voronoi图的拓扑数据。基于这种拓扑结构，构造了一个符号距离场来隐式地表示多孔模型，为孔隙形态的表达提供了特殊的灵活性。其次，利用描述自然生长过程的生长函数，在多孔模型中实现孔隙形态的平滑过渡，实现对过渡的精确控制，便于分层和梯度多孔结构的设计。最后，通过引入径向基函数（RBF）加权插值方法，将过渡区域推广到整个区域，从而实现复杂的梯度设计和仿生仿真。该研究为复杂拓扑的建模提供了一种新的解决方案。

{"title":"Implicit complex topology modeling design method based on Voronoi","authors":"Chongshui Liu, Quan Qi","doi":"10.1016/j.cad.2025.103910","DOIUrl":"10.1016/j.cad.2025.103910","url":null,"abstract":"<div><div>Porous structures with complex topologies have long been at the forefront of research in the field of additive manufacturing. However, existing methods exhibit limited capability in representing the morphology of pores in random porous structures, lack precision in gradient design, and struggle to accurately simulate the stochastic porous features found in nature. This paper proposes a novel modeling approach for random porous structures. Firstly, the method utilizes Voronoi tessellation to obtain the topological data of three-dimensional Voronoi diagrams. Based on this topology, a signed distance field is constructed to implicitly represent the porous model, providing exceptional flexibility in the expression of pore morphologies. Secondly, growth functions that describe natural growth processes are employed to achieve smooth transitions in pore morphology within the porous model, enabling precise control over transitions and facilitating the design of hierarchical and gradient porous structures. Finally, by introducing a radial basis function (RBF) weighted interpolation method, the transition region is generalized to the entire domain, allowing for complex gradient designs and biomimetic simulations. This study offers a new solution for the modeling of complex topologies.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"188 ","pages":"Article 103910"},"PeriodicalIF":3.0,"publicationDate":"2025-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144670495","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Reinforcement learning-based parametric CAD models reconstruction from 2D orthographic drawings 基于强化学习的二维正射影图参数化CAD模型重建

IF 3 3区计算机科学 Q2 COMPUTER SCIENCE, SOFTWARE ENGINEERING

Computer-Aided Design

Pub Date : 2025-07-16 DOI: 10.1016/j.cad.2025.103925

Chao Zhang , Arnaud Polette , Romain Pinquié , Mirai Iida , Henri De Charnace , Jean-Philippe Pernot

This paper introduces a reinforcement learning-based approach for reconstructing 3D parametric CAD models from 2D orthographic drawings. First, the 2D drawings are parsed to extract their constituent vertices and edges. These entities are subsequently converted into a newly defined loop-path representation, generating a list of loop-path pairs along with their associated parameters and candidates for the reconstruction process. The core of the approach is a DQN-based agent trained to select the sequences of loop-path pairs, which are then used to reconstruct the parametric CAD models in any CAD modeler. A parallel environment leveraging a neural network is proposed to accelerate the training process and eliminate the need for calls to an external CAD modeler to compute the rewards, which would otherwise break the training loop. The proposed approach reconstructs 3D parametric CAD models in less than a second, and it outperforms existing methods against traditional metrics on two datasets. The reconstructed CAD models are fully editable and can be easily modified for downstream applications. While the loop-path representation supports extrusion, revolution and sweep operations, experimental results on the two selected datasets highlight the superiority of the RL-based approach in handling sketch-extrude modeling operations.

本文介绍了一种基于强化学习的方法，用于从二维正射影图重构三维参数化CAD模型。首先，对二维图形进行解析，提取其组成顶点和边缘；这些实体随后被转换为新定义的环路路径表示，生成环路路径对列表及其相关参数和用于重建过程的候选项。该方法的核心是一个基于dqn的智能体，它被训练来选择环路路径对的序列，然后用于在任何CAD建模器中重建参数化CAD模型。提出了一种利用神经网络的并行环境来加速训练过程，并消除了调用外部CAD建模器来计算奖励的需要，否则会破坏训练循环。该方法在不到1秒的时间内重建三维参数化CAD模型，并且在两个数据集上优于现有的传统度量方法。重建的CAD模型是完全可编辑的，可以很容易地修改下游应用程序。虽然环路表示支持挤压、旋转和扫描操作，但在两个选定的数据集上的实验结果突出了基于rl的方法在处理草图挤压建模操作方面的优势。

{"title":"Reinforcement learning-based parametric CAD models reconstruction from 2D orthographic drawings","authors":"Chao Zhang , Arnaud Polette , Romain Pinquié , Mirai Iida , Henri De Charnace , Jean-Philippe Pernot","doi":"10.1016/j.cad.2025.103925","DOIUrl":"10.1016/j.cad.2025.103925","url":null,"abstract":"<div><div>This paper introduces a reinforcement learning-based approach for reconstructing 3D parametric CAD models from 2D orthographic drawings. First, the 2D drawings are parsed to extract their constituent vertices and edges. These entities are subsequently converted into a newly defined loop-path representation, generating a list of loop-path pairs along with their associated parameters and candidates for the reconstruction process. The core of the approach is a DQN-based agent trained to select the sequences of loop-path pairs, which are then used to reconstruct the parametric CAD models in any CAD modeler. A parallel environment leveraging a neural network is proposed to accelerate the training process and eliminate the need for calls to an external CAD modeler to compute the rewards, which would otherwise break the training loop. The proposed approach reconstructs 3D parametric CAD models in less than a second, and it outperforms existing methods against traditional metrics on two datasets. The reconstructed CAD models are fully editable and can be easily modified for downstream applications. While the loop-path representation supports extrusion, revolution and sweep operations, experimental results on the two selected datasets highlight the superiority of the RL-based approach in handling sketch-extrude modeling operations.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"188 ","pages":"Article 103925"},"PeriodicalIF":3.0,"publicationDate":"2025-07-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144655030","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

CADInstruct: A multimodal dataset for natural language-guided CAD program synthesis 用于自然语言引导的CAD程序合成的多模态数据集

IF 3 3区计算机科学 Q2 COMPUTER SCIENCE, SOFTWARE ENGINEERING

Computer-Aided Design

Pub Date : 2025-07-12 DOI: 10.1016/j.cad.2025.103926

Chaofan Lv, Jinsong Bao

While large language models (LLMs) have demonstrated remarkable success in general-purpose code generation, their application in computer-aided design (CAD) program synthesis remains constrained by the scarcity of high-quality natural language-annotated datasets. To address this challenge, we propose CADInstruct, a novel approach aimed at constructing a multimodal CAD instruction dataset to enhance the CAD program synthesis capabilities of LLMs. First, we introduce a parametric modification module for modeling sequences, which extracts geometric constraints and critical dimensions from sketches, transforming CAD construction sequences into design-intent-oriented instructions. Second, we incorporate a shape semantic recognition module that leverages model names and visually enriched rendered views to generate precise shape descriptions using multimodal large models, enabling accurate semantic representation of complex geometries. Lastly, a modeling instruction semantic alignment module utilizes the extracted shape descriptions and modeling instructions to generate hierarchical natural language descriptions, encompassing geometric forms and detailed modeling steps, ensuring consistency between textual descriptions and CAD instructions. We fine-tuned the Qwen2.5-Coder-7B model using the CADInstruct dataset to evaluate the effectiveness of this framework. Experimental results demonstrated its capability to significantly enhance CAD program synthesis. The code and dataset will be made publicly available at https://github.com/dxlcf/CADInstruct.

虽然大型语言模型（llm）在通用代码生成方面取得了显著的成功，但它们在计算机辅助设计（CAD）程序合成中的应用仍然受到高质量自然语言注释数据集的缺乏的限制。为了解决这一挑战，我们提出了caddirective，一种旨在构建多模态CAD指令数据集的新方法，以增强llm的CAD程序合成能力。首先，我们引入了建模序列的参数化修改模块，该模块从草图中提取几何约束和关键尺寸，将CAD构造序列转换为面向设计意图的指令。其次，我们整合了一个形状语义识别模块，该模块利用模型名称和视觉上丰富的渲染视图，使用多模态大模型生成精确的形状描述，从而实现复杂几何形状的准确语义表示。最后，建模指令语义对齐模块利用提取的形状描述和建模指令生成分层自然语言描述，包括几何形式和详细的建模步骤，确保文本描述与CAD指令之间的一致性。我们使用caddirective数据集对qwen2.5 - code - 7b模型进行了微调，以评估该框架的有效性。实验结果表明，该方法能够显著提高CAD程序的综合能力。代码和数据集将在https://github.com/dxlcf/CADInstruct上公开提供。

{"title":"CADInstruct: A multimodal dataset for natural language-guided CAD program synthesis","authors":"Chaofan Lv, Jinsong Bao","doi":"10.1016/j.cad.2025.103926","DOIUrl":"10.1016/j.cad.2025.103926","url":null,"abstract":"<div><div>While large language models (LLMs) have demonstrated remarkable success in general-purpose code generation, their application in computer-aided design (CAD) program synthesis remains constrained by the scarcity of high-quality natural language-annotated datasets. To address this challenge, we propose CADInstruct, a novel approach aimed at constructing a multimodal CAD instruction dataset to enhance the CAD program synthesis capabilities of LLMs. First, we introduce a parametric modification module for modeling sequences, which extracts geometric constraints and critical dimensions from sketches, transforming CAD construction sequences into design-intent-oriented instructions. Second, we incorporate a shape semantic recognition module that leverages model names and visually enriched rendered views to generate precise shape descriptions using multimodal large models, enabling accurate semantic representation of complex geometries. Lastly, a modeling instruction semantic alignment module utilizes the extracted shape descriptions and modeling instructions to generate hierarchical natural language descriptions, encompassing geometric forms and detailed modeling steps, ensuring consistency between textual descriptions and CAD instructions. We fine-tuned the Qwen2.5-Coder-7B model using the CADInstruct dataset to evaluate the effectiveness of this framework. Experimental results demonstrated its capability to significantly enhance CAD program synthesis. The code and dataset will be made publicly available at <span><span>https://github.com/dxlcf/CADInstruct</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"188 ","pages":"Article 103926"},"PeriodicalIF":3.0,"publicationDate":"2025-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144679853","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Functional map-based reflection intrinsic symmetry detection and symmetrization for 2D deformable shapes 二维可变形形状的基于功能映射反射的内禀对称性检测与对称化

IF 3 3区计算机科学 Q2 COMPUTER SCIENCE, SOFTWARE ENGINEERING

Computer-Aided Design

Pub Date : 2025-07-10 DOI: 10.1016/j.cad.2025.103908

Shengjun Liu , Zi Teng , Haibo Wang , Xinru Liu

Symmetry is widely prevalent in both natural phenomena and man-made objects. Detecting and enhancing the symmetry of shapes is crucial in fields like art and engineering. However, symmetry detection and shape symmetrization based on it for 2D deformable shapes have long been a challenge. In this paper, we propose a 2D intrinsic symmetry detection method based on functional maps. By leveraging constraints from feature-symmetric point pairs and functional maps framework, our approach formulates an optimization problem for detecting intrinsic symmetry in 2D deformable shapes. We employ spectral upsampling techniques, iteratively optimizing the functional map matrix and symmetric point-to-point mapping matrix in both frequency and spatial domains. Then we perform rapid symmetrization guided by the extracted backbone and shape symmetry. The backbone is detected by mapping skeletons, and projecting it onto a straight line segment drives automatic mesh deformation to symmetrize the 2D shape. We have tested our method on a variety of 2D shapes, and the results have demonstrated its effectiveness.

对称在自然现象和人造物体中普遍存在。检测和增强形状的对称性在艺术和工程等领域至关重要。然而，基于该方法的二维可变形形状的对称检测和形状对称一直是一个难题。本文提出了一种基于功能映射的二维本征对称性检测方法。通过利用特征对称点对和功能映射框架的约束，我们的方法制定了一个优化问题，用于检测二维可变形形状的内在对称性。我们采用频谱上采样技术，在频率域和空间域迭代优化功能映射矩阵和对称点对点映射矩阵。然后根据提取的骨架和形状对称进行快速对称。通过映射骨架来检测骨架，并将其投影到直线段上，驱动自动网格变形以使二维形状对称。我们已经在各种二维形状上测试了我们的方法，结果证明了它的有效性。

{"title":"Functional map-based reflection intrinsic symmetry detection and symmetrization for 2D deformable shapes","authors":"Shengjun Liu , Zi Teng , Haibo Wang , Xinru Liu","doi":"10.1016/j.cad.2025.103908","DOIUrl":"10.1016/j.cad.2025.103908","url":null,"abstract":"<div><div>Symmetry is widely prevalent in both natural phenomena and man-made objects. Detecting and enhancing the symmetry of shapes is crucial in fields like art and engineering. However, symmetry detection and shape symmetrization based on it for 2D deformable shapes have long been a challenge. In this paper, we propose a 2D intrinsic symmetry detection method based on functional maps. By leveraging constraints from feature-symmetric point pairs and functional maps framework, our approach formulates an optimization problem for detecting intrinsic symmetry in 2D deformable shapes. We employ spectral upsampling techniques, iteratively optimizing the functional map matrix and symmetric point-to-point mapping matrix in both frequency and spatial domains. Then we perform rapid symmetrization guided by the extracted backbone and shape symmetry. The backbone is detected by mapping skeletons, and projecting it onto a straight line segment drives automatic mesh deformation to symmetrize the 2D shape. We have tested our method on a variety of 2D shapes, and the results have demonstrated its effectiveness.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"188 ","pages":"Article 103908"},"PeriodicalIF":3.0,"publicationDate":"2025-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144655029","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

SDF-CWF: Consolidating Weak Features in High-Quality Mesh Extraction from Signed Distance Functions SDF-CWF：从带符号距离函数中提取高质量网格的弱特征整合

IF 3 3区计算机科学 Q2 COMPUTER SCIENCE, SOFTWARE ENGINEERING

Computer-Aided Design

Pub Date : 2025-07-08 DOI: 10.1016/j.cad.2025.103912

Longdu Liu (Researcher) , Hao Yu , Shiqing Xin , Shuangmin Chen , Hongwei Lin , Wenping Wang , Changhe Tu

With advancements in geometric deep learning techniques, neural signed distance functions (SDFs) have gained popularity for their flexibility. Recent studies show that neural SDFs can retain geometric details and encode sharp features. However, during the mesh extraction stage, methods like marching cubes may degrade these geometric details and sharp features, thus compromising the expressiveness of neural SDFs.

In this paper, we aim to develop a general-purpose mesh extraction method for both freeform and CAD models, assuming the availability of a SDF. Our goal is to produce a well-triangulated, resolution-adjustable mesh surface that preserves rich geometric details and distinct feature lines. Our approach is inspired by Centroidal Voronoi Tessellation (CVT) but introduces two key modifications. First, we extend CVT computation to implicit representations, where explicit surface decomposition is not available. Second, we propose a measure for estimating the likelihood that a point lies on feature lines, enabling the extraction of feature-aligned triangle meshes using power diagrams (with site weights positively correlated to the likelihood values). Comprehensive comparisons with state-of-the-art methods demonstrate the superiority of our approach in both feature alignment and triangulation quality.

随着几何深度学习技术的进步，神经符号距离函数（sdf）因其灵活性而受到欢迎。最近的研究表明，神经sdf可以保留几何细节并编码尖锐特征。然而，在网格提取阶段，像行进立方体这样的方法可能会降低这些几何细节和尖锐特征，从而影响神经sdf的表达性。在本文中，我们的目标是为自由形状和CAD模型开发一种通用的网格提取方法，假设SDF的可用性。我们的目标是产生一个良好的三角化，分辨率可调的网格表面，保留丰富的几何细节和鲜明的特征线。我们的方法受到质心Voronoi镶嵌（CVT）的启发，但引入了两个关键的修改。首先，我们将CVT计算扩展到隐式表示，其中显式表面分解是不可用的。其次，我们提出了一种估计点位于特征线上的可能性的方法，从而可以使用功率图（站点权重与似然值正相关）提取特征对齐的三角形网格。与最先进的方法进行综合比较，证明了我们的方法在特征对齐和三角测量质量方面的优越性。

{"title":"SDF-CWF: Consolidating Weak Features in High-Quality Mesh Extraction from Signed Distance Functions","authors":"Longdu Liu (Researcher) , Hao Yu , Shiqing Xin , Shuangmin Chen , Hongwei Lin , Wenping Wang , Changhe Tu","doi":"10.1016/j.cad.2025.103912","DOIUrl":"10.1016/j.cad.2025.103912","url":null,"abstract":"<div><div>With advancements in geometric deep learning techniques, neural signed distance functions (SDFs) have gained popularity for their flexibility. Recent studies show that neural SDFs can retain geometric details and encode sharp features. However, during the mesh extraction stage, methods like marching cubes may degrade these geometric details and sharp features, thus compromising the expressiveness of neural SDFs.</div><div>In this paper, we aim to develop a general-purpose mesh extraction method for both freeform and CAD models, assuming the availability of a SDF. Our goal is to produce a well-triangulated, resolution-adjustable mesh surface that preserves rich geometric details and distinct feature lines. Our approach is inspired by Centroidal Voronoi Tessellation (CVT) but introduces two key modifications. First, we extend CVT computation to implicit representations, where explicit surface decomposition is not available. Second, we propose a measure for estimating the likelihood that a point lies on feature lines, enabling the extraction of feature-aligned triangle meshes using power diagrams (with site weights positively correlated to the likelihood values). Comprehensive comparisons with state-of-the-art methods demonstrate the superiority of our approach in both feature alignment and triangulation quality.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"188 ","pages":"Article 103912"},"PeriodicalIF":3.0,"publicationDate":"2025-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144596859","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A method of surface mesh generation for industrial CAD models by constructing conforming discrete representation 一种通过构造符合离散表示的工业CAD模型曲面网格生成方法

IF 3 3区计算机科学 Q2 COMPUTER SCIENCE, SOFTWARE ENGINEERING

Computer-Aided Design

Pub Date : 2025-06-30 DOI: 10.1016/j.cad.2025.103914

Xin Yang , Fei Yu , Yan Zhou , Pengchao Zhou , G.Z. Zhao , Z.Q. Guan

In this paper, we present a new approach for surface mesh generation of industrial CAD models by establishing a conforming discrete representation. Firstly, we develop a two-step strategy, A-Boolean, which combines the aligned surface meshing (ASM) and Boolean operations to generate conforming meshes for industrial CAD models with defects including gaps, contacts, overlaps and self-intersections. During the mesh generation, ASM can properly solve tiny gaps, contacts and intersections by given a small threshold. Then, the remaining defects can be repaired by the Boolean operations and avoid numerical errors. For each surface, the approximation error between the discrete model and original geometry is well controlled by appending orthogonal non-uniform grid points. Secondly, an additional topology reconstruction process is introduced to transform conforming meshes into manifold partitions, which enhance the reliability and versatility of the following meshing algorithm. During the construction of discrete representation, geometric issues that affect the reliability and quality of mesh generation can be effectively solved. Finally, high-quality meshes can be generated by a remeshing algorithm based on surface parameterization. Several examples are presented with comparisons against open-source and commercial software, demonstrating the validity of the proposed method.

本文提出了一种工业CAD模型曲面网格生成的新方法，即建立符合要求的离散表示。首先，我们开发了一种两步策略，a -布尔，它结合了对准曲面网格（ASM）和布尔运算，为具有间隙、接触、重叠和自交等缺陷的工业CAD模型生成符合的网格。在网格生成过程中，ASM可以通过给定较小的阈值，较好地求解微小的间隙、接触和交集。然后通过布尔运算修复剩余缺陷，避免数值误差。对于每个曲面，通过附加正交非均匀网格点，很好地控制了离散模型与原始几何形状之间的逼近误差。其次，引入额外的拓扑重构过程，将符合条件的网格转化为流形分区，提高了后续网格划分算法的可靠性和通用性；在离散表示的构建过程中，可以有效地解决影响网格生成可靠性和质量的几何问题。最后，采用基于曲面参数化的重划分算法生成高质量的网格。通过与开源软件和商业软件的比较，验证了该方法的有效性。

{"title":"A method of surface mesh generation for industrial CAD models by constructing conforming discrete representation","authors":"Xin Yang , Fei Yu , Yan Zhou , Pengchao Zhou , G.Z. Zhao , Z.Q. Guan","doi":"10.1016/j.cad.2025.103914","DOIUrl":"10.1016/j.cad.2025.103914","url":null,"abstract":"<div><div>In this paper, we present a new approach for surface mesh generation of industrial CAD models by establishing a conforming discrete representation. Firstly, we develop a two-step strategy, A-Boolean, which combines the aligned surface meshing (ASM) and Boolean operations to generate conforming meshes for industrial CAD models with defects including gaps, contacts, overlaps and self-intersections. During the mesh generation, ASM can properly solve tiny gaps, contacts and intersections by given a small threshold. Then, the remaining defects can be repaired by the Boolean operations and avoid numerical errors. For each surface, the approximation error between the discrete model and original geometry is well controlled by appending orthogonal non-uniform grid points. Secondly, an additional topology reconstruction process is introduced to transform conforming meshes into manifold partitions, which enhance the reliability and versatility of the following meshing algorithm. During the construction of discrete representation, geometric issues that affect the reliability and quality of mesh generation can be effectively solved. Finally, high-quality meshes can be generated by a remeshing algorithm based on surface parameterization. Several examples are presented with comparisons against open-source and commercial software, demonstrating the validity of the proposed method.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"188 ","pages":"Article 103914"},"PeriodicalIF":3.0,"publicationDate":"2025-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144556890","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

High-resolution fiber-level simulation of knitted patterns 高分辨率纤维级针织图案模拟

IF 3 3区计算机科学 Q2 COMPUTER SCIENCE, SOFTWARE ENGINEERING

Computer-Aided Design

Pub Date : 2025-06-30 DOI: 10.1016/j.cad.2025.103913

Xin Yang , Cheng Lu , Huiqi Shao , Guangwei Shao , Jinhua Jiang , Siyi Bi , Nanliang Chen

Knitted fabrics, characterized by intricate patterns, vibrant colors, and soft tactile properties, have long served as a source of inspiration in textile design. Leveraging digital technology to translate these design concepts into realistic models, this paper proposes a fiber-level 3D simulation framework for complex knitted structures, inspired by the digital element methodology. In this approach, yarns are discretized into fiber assemblies represented by sequences of control points. Improved beam elements connect adjacent points to model bending behavior, while rod elements simulate inter-fiber interactions. To improve structural controllability, dynamic boundary conditions and variable driving forces are introduced, enabling accurate capture of both global and local deformations. An efficient Array operation is developed to support scalable generation of fabric patterns under a modified periodic boundary condition. Experimental evaluations demonstrate that the proposed method achieves visually and structurally accurate simulations within a limited number of iterations. Comparative analysis with real fabric samples validates the effectiveness and fidelity of the simulation framework, making it suitable for applications in virtual textile design and performance prediction.

针织面料以其复杂的图案、鲜艳的色彩和柔软的触感为特点，长期以来一直是纺织品设计的灵感来源。利用数字技术将这些设计概念转化为现实模型，本文提出了一种受数字元素方法启发的复杂针织结构的纤维级3D模拟框架。在这种方法中，纱线被离散成由控制点序列表示的纤维组合。改进的梁单元连接相邻点来模拟弯曲行为，而杆单元模拟纤维间的相互作用。为了提高结构的可控性，引入了动态边界条件和可变驱动力，从而能够准确捕获全局和局部变形。在改进的周期边界条件下，开发了一种高效的Array操作，支持可扩展的织物图案生成。实验结果表明，该方法在有限的迭代次数内实现了视觉上和结构上的精确模拟。通过与真实织物样品的对比分析，验证了仿真框架的有效性和保真度，使其适合于虚拟纺织品设计和性能预测的应用。

{"title":"High-resolution fiber-level simulation of knitted patterns","authors":"Xin Yang , Cheng Lu , Huiqi Shao , Guangwei Shao , Jinhua Jiang , Siyi Bi , Nanliang Chen","doi":"10.1016/j.cad.2025.103913","DOIUrl":"10.1016/j.cad.2025.103913","url":null,"abstract":"<div><div>Knitted fabrics, characterized by intricate patterns, vibrant colors, and soft tactile properties, have long served as a source of inspiration in textile design. Leveraging digital technology to translate these design concepts into realistic models, this paper proposes a fiber-level 3D simulation framework for complex knitted structures, inspired by the digital element methodology. In this approach, yarns are discretized into fiber assemblies represented by sequences of control points. Improved beam elements connect adjacent points to model bending behavior, while rod elements simulate inter-fiber interactions. To improve structural controllability, dynamic boundary conditions and variable driving forces are introduced, enabling accurate capture of both global and local deformations. An efficient Array operation is developed to support scalable generation of fabric patterns under a modified periodic boundary condition. Experimental evaluations demonstrate that the proposed method achieves visually and structurally accurate simulations within a limited number of iterations. Comparative analysis with real fabric samples validates the effectiveness and fidelity of the simulation framework, making it suitable for applications in virtual textile design and performance prediction.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"188 ","pages":"Article 103913"},"PeriodicalIF":3.0,"publicationDate":"2025-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144571347","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0