基于Weisfeiler-Lehman核增强的无监督BIM构件检索方法

doi:10.11996/JG.j.2095-302X.2025051123

图学学报 ›› 2025, Vol. 46 ›› Issue (5): 1123-1133.DOI: 10.11996/JG.j.2095-302X.2025051123

基于Weisfeiler-Lehman核增强的无监督BIM构件检索方法

扈慧强¹(), 贺长雁², 刘小军³(), 贾金原¹, 高路⁴

¹ 同济大学软件学院，上海 201804
² 嘉兴大学师范学院，浙江嘉兴 314001
³ 嘉兴南湖学院信息工程学院，浙江嘉兴 314001
⁴ 因筑(苏州)建筑科技有限公司，江苏苏州 215131

收稿日期:2024-12-25 接受日期:2025-03-11 出版日期:2025-10-30 发布日期:2025-09-10
通讯作者:刘小军(1981-)，男，副教授，博士。主要研究方向为计算机视觉与数字孪生等。E-mail：xjliu1204@126.com
第一作者:扈慧强(1983-)，男，博士研究生。主要研究方向为计算机图形学与建筑数据分析等。E-mail：276349967@qq.com
基金资助:
浙江省公益技术应用研究计划一般项目(LGG22F020037);国家自然科学基金面上项目(6207071897);国家自然科学区域联合基金重点项目(U19A2063)

Towards unsupervised BIM product retrieval: a Weisfeiler-Lehman kernel enhanced approach

HU Huiqiang¹(), HE Changyan², LIU Xiaojun³(), JIA Jinyuan¹, GAO Lu⁴

¹ School of Software Engineering, Tongji University, Shanghai 201804, China
² Normal College, Jiaxing University, Jiaxing Zhejiang 314001, China
³ College of Information Engineering, Jiaxing Nanhu University, Jiaxing Zhejiang 304001, China
⁴ Yinzhu (Suzhou) Building Technology Co., LTD, Suzhou Jiangsu 215131, China

Received:2024-12-25 Accepted:2025-03-11 Published:2025-10-30 Online:2025-09-10
First author：HU Huiqiang (1983-), PhD candidate. His main research interests cover computer graphics and building data analysis, etc. E-mail：276349967@qq.com
Supported by:
Zhejiang Province Basic Public Welfare Research Program Project of China(LGG22F020037);National Natural Science Foundation of China General Program(6207071897);National Natural Science Foundation of China Key Program of the Regional Joint Fund(U19A2063)

摘要/Abstract

摘要：

针对建筑行业对构件检索的迫切需求，提出一种面向工业基础类(IFC)数据特点的无监督建筑信息模型(BIM)构件检索方法。充分利用IFC标准中的语义和几何信息，构建了构件属性图(PAG)作为构件特征，并结合PAG的多属性通道，提出基于Weisfeiler-Lehman (WL)核增强的PAG同构预测思路，以实现BIM构件检索。该方法支持以2个IFC文档作为输入，其中文档A代表待检索的构件，文档B则为构件库，最终返回B中与A相似的构件。主要贡献在于：①提出一个无须数据预处理且兼容语义信息的BIM构件检索框架；②引入BIM构件的PAG特征构建方法，同时，给出WL图核增强的PAG同构预测方法；③设计了无监督的收敛性判断策略，通过分析预测结果中的属性差异及时判断算法的收敛状态。实验结果表明，该方法的PAG同构测试能在不超过3次迭代的情况下收敛，实验环境下BIM构件的同构测试不超过1 s，构件搜索的平均准确率达95%。

关键词: 建筑构件, Weisfeiler-Lehman核, 检索, 无监督, 图同构

Abstract:

To meet the urgent need for building elements retrieval in the construction industry, an unsupervised building information modeling (BIM) product retrieval method tailored to the characteristics of industry foundation classes (IFC) data was proposed. The method fully exploited the semantic and geometric information from the IFC standard to construct a product attributed graph (PAG) as the product feature. By leveraging the multi-attribute channels of PAG, a PAG isomorphism prediction approach, enhanced by the Weisfeiler-Lehman (WL) kernel, was proposed to achieve BIM product retrieval. The proposed method accepted two IFC documents as input: Document A, representing the target product to be retrieved, and Document B, serving as the product library. Our method ultimately returned products from Document B similar to the target product in Document A. The principal contributions were threefold: ①The proposal of a BIM product retrieval framework that circumvented the need for data preprocessing while maintaining semantic integrity. ②The development of PAG feature extraction for BIM product and enhanced PAG isomorphism prediction method with augmented WL graph kernels. ③The design of an unsupervised convergence assessment strategy in which the convergence status was timely determined by analyzing the attribute differences between the attributes from source and those predicted. Empirical findings indicated that the PAG isomorphism testing of our methodology achieved convergence within a maximum of three iterations. Under the experimental conditions, the isomorphism testing of BIM products required no longer than 1 second, with an average accuracy rate of 95% in products retrieval.

Key words: building product, Weisfeiler-Lehman kernel, retrieval, unsupervised, graph isomorphism

中图分类号:

TP391.3
TU17

扈慧强, 贺长雁, 刘小军, 贾金原, 高路. 基于Weisfeiler-Lehman核增强的无监督BIM构件检索方法[J]. 图学学报, 2025, 46(5): 1123-1133.

HU Huiqiang, HE Changyan, LIU Xiaojun, JIA Jinyuan, GAO Lu. Towards unsupervised BIM product retrieval: a Weisfeiler-Lehman kernel enhanced approach[J]. Journal of Graphics, 2025, 46(5): 1123-1133.

图/表 14

图1 BIM构件引用部件的IFC文档示例

Fig. 1 An example for BIM product and its referenced components in IFC document

图2 1-WL的一次迭代过程

Fig. 2 An iterator process of 1-WL

图3 BIM构件属性图

Fig. 3 An example of PAG

图4 PAG的属性压缩操作

Fig. 4 The compress operation of PAG attributes

图5 基于PAG的WL核增强算法迭代过程(单个PAG的角度)

Fig. 5 An iterator of our WL kernel enhanced algorithm (from a single PAG’s view)

图6 本文方法优于其他同类方法的典型案例[27]

Fig. 6 A typical case demonstrating the superiority of our method over others within the same series[27]

表1 实验数据参数

Table 1 Experimental data parameters

构件类型	P1	P2	P3	P4	P5	P6	总计
IfcBuildingElementProxy	2	1			9	208	220
IfcDoor	10			11		102
IfcEnergyConversionDevice	137						137
IfcFlowFitting	2						2
IfcFlowTerminal	68	3		37		57	165
IfcFurnishingElement		27		79	7	434	684
IfcRailing	22	7		35		221	287
IfcStair		3					3
IfcStairFlight				20			20
IfcTransportElement		2					2
IfcWindow						53	53
总计	241	43	39	182	16	1 175	1 696

图7 数据样例(rac_advanced_sample_project)的可视化效果

Fig. 7 The visualization of sample data (rac_advanced_sample_project)

图8 数据集中对应的PAG节点分布

Fig. 8 PAG nodes distribution within our dataset

图9 算法迭代次数比较结果

Fig. 9 Comparison analysis of iteration counts

图10 算法的时间损耗对比结果

Fig. 10 Comparison analysis of time consumption

图11 BIM构件检索准确率对比结果

Fig. 11 Comparison analysis of retrieval accuracy of BIM products

图12 拓扑相同但PAG属性不同的BIM构件实例

Fig. 12 BIM products with identical topology but distinct PAG attributes

图13 3种实例中部件间的距离分布

Fig. 13 The distances distribution of components from three types of instances ((a) IfcBuildingElementProxy; (b) IfcFurnishingElement; (c) IfcRailing)

参考文献 28

[1]	CAO J P, BUCHER D F, HALL D M, et al. A graph-based approach for module library development in industrialized construction[J]. Computers in Industry, 2022, 139: 103659.
[2]	LI N X, LI Q, LIU Y S, et al. BIMSeek++: retrieving BIM components using similarity measurement of attributes[J]. Computers in Industry, 2020, 116: 103186.
[3]	ZHENG J W, FISCHER M. Dynamic prompt-based virtual assistant framework for BIM information search[J]. Automation in Construction, 2023, 155: 105067.
[4]	刘小军, 贺长雁, 刘畅, 等. 基于结构匹配的建筑信息模型构件快速对齐方法[J]. 系统仿真学报, 2021, 33(7): 1626-1637. DOI
	LIU X J, HE C Y, LIU C, et al. Fast alignment of BIM products based on structure matching[J]. Journal of System Simulation, 2021, 33(7): 1626-1637 (in Chinese). DOI
[5]	SHAO Y X, LI H Z, GU X Z, et al. Distributed graph neural network training: a survey[J]. ACM Computing Surveys, 2024, 56(8): 191.
[6]	姜柳, 史健勇, 付功义, 等. 基于BIM和深度学习的建筑平面凹凸不规则识别[J]. 图学学报, 2022, 43(3): 522-529.
	JIANG L, SHI J Y, FU G Y, et al. Identification of the plane irregularity of structures based on BIM and deep learning[J]. Journal of Graphics, 2022, 43(3): 522-529 (in Chinese).
[7]	DAI C Y, CHENG K, LIANG B X, et al. Digital twin modeling method based on IFC standards for building construction processes[J]. Frontiers in Energy Research, 2024, 12: 1334192.
[8]	FUHLBRÜCK F, KÖBLER J, PONOMARENKO I, et al. The Weisfeiler-Leman algorithm and recognition of graph properties[J]. Theoretical Computer Science, 2021, 895: 96-114.
[9]	DU X Z, GU Y X, YANG N, et al. IFC file content compression based on reference relationships[J]. Journal of Computing in Civil Engineering, 2020, 34(3): 04020012.
[10]	刘小军, 贾金原. 面向手机网页的大规模WebBIM场景轻量级实时漫游算法[J]. 中国科学: 信息科学, 2018, 48(3): 274-292.
	LIU X J, JIA J Y. Mobile web-based lightweight and real-time roaming algorithm for large-scale WebBIM scenes[J]. SCIENTIA SINICA Informationis, 2018, 48(3): 274-292 (in Chinese).
[11]	DU X Z, DUAN Y R. Submodel comparison method based on IFC file content[J]. Journal of Computing in Civil Engineering, 2022, 36(4): 04022014.
[12]	王煜, 邓晖, 李晓瑶, 等. 自然语言处理技术在建筑工程中的应用研究综述[J]. 图学学报, 2020, 41(4): 501-511. DOI
	WANG Y, DENG H, LI X Y, et al. A review of natural language processing application in construction engineering[J]. Journal of Graphics, 2020, 41(4): 501-511 (in Chinese). DOI
[13]	WANG N, ISSA R R A, ANUMBA C J. NLP-based query-answering system for information extraction from building information models[J]. Journal of Computing in Civil Engineering, 2022, 36(3): 04022004.
[14]	WANG Z J, SACKS R, YEUNG T. Exploring graph neural networks for semantic enrichment: room type classification[J]. Automation in Construction, 2022, 134: 104039.
[15]	RODRIGUES R S V, MORGADO J F M, GOMES A J P. Part-based mesh segmentation: a survey[J]. Computer Graphics Forum, 2018, 37(6): 235-274.
[16]	KIM J, MO K, SUNG M, et al. Seg&struct: the interplay between part segmentation and structure inference for 3D shape parsing[C]// 2023 IEEE/CVF Winter Conference on Applications of Computer Vision. New York: IEEE Press, 2023: 1226-1235.
[17]	NIEPERT M, AHMED M, KUTZKOV K. Learning convolutional neural networks for graphs[EB/OL]. [2024- 10-24]. https://dblp.uni-trier.de/db/conf/icml/icml2016.html#NiepertAK16.
[18]	WEI X, YU R X, SUN J. View-GCN: view-based graph convolutional network for 3D shape analysis[C]// 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2020: 1850-1859.
[19]	WANG Z M, CHEN J, CHEN H P. EGAT: edge-featured graph attention network[C]// The 30th International Conference on Artificial Neural Networks. Cham: Springer, 2021: 253-264.
[20]	PEI H B, WEI B Z, CHANG K C C, et al. Geom-GCN: geometric graph convolutional networks[EB/OL]. [2024- 10-24]. https://dblp.uni-trier.de/db/conf/iclr/iclr2020.html#PeiWCLY20.
[21]	HAN J Q, CEN J C, WU L M, et al. A survey of geometric graph neural networks: data structures, models and applications[EB/OL]. [2024-10-24]. https://arxiv.org/abs/2403.00485.
[22]	ARVIND V, FUHLBRÜCK F, KÖBLER J, et al. On Weisfeiler-Leman invariance: subgraph counts and related graph properties[J]. Journal of Computer and System Sciences, 2020, 113: 42-59.
[23]	ZHANG B H, GAI J C, DU Y H, et al. Beyond Weisfeiler-Lehman: a quantitative framework for GNN expressiveness[EB/OL]. [2024-10-24]. https://dblp.uni-trier.de/db/conf/iclr/iclr2024.html#ZhangGDY0024.
[24]	CHANG L J, FENG X, YAO K, et al. Accelerating graph similarity search via efficient GED computation[J]. IEEE Transactions on Knowledge and Data Engineering, 2023, 35(5): 4485-4498.
[25]	ZHOU X P, MA C X, WANG M M, et al. BIM product recommendation for intelligent design using style learning[J]. Journal of Building Engineering, 2023, 73: 106701.
[26]	JOHNSON S J, MURTY M R, NAVAKANTH I. A detailed review on word embedding techniques with emphasis on word2vec[J]. Multimedia Tools and Applications, 2024, 83(13): 37979-38007.
[27]	SATO R. A survey on the expressive power of graph neural networks[EB/OL]. [2024-10-24]. https://arxiv.org/abs/2003.04078.
[28]	LIU X J, XIE N, TANG K, et al. Lightweighting for Web3D visualization of large-scale BIM scenes in real-time[J]. Graphical Models, 2016, 88: 40-56.

基于Weisfeiler-Lehman核增强的无监督BIM构件检索方法

Towards unsupervised BIM product retrieval: a Weisfeiler-Lehman kernel enhanced approach

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 14

参考文献 28

相关文章 15

编辑推荐

Metrics

本文评价

[1]	凌非, 余京涛, 朱哲燕, 罗剑, 朱继祥, 陈先客, 董建锋. 基于对比学习的数据高效视频检索[J]. 图学学报, 2025, 46(3): 491-501.
[2]	郑洪岩, 王慧, 刘昊, 张志平, 杨晓娟, 孙涛. 基于隐式知识增强的KB-VQA知识检索策略研究[J]. 图学学报, 2024, 45(6): 1231-1242.
[3]	刘以, 邱军海, 张嘉星, 张小峰, 王桦, 张彩明. 基于权衡因子和多维空间度量的高鲁棒性图像分割算法[J]. 图学学报, 2024, 45(3): 482-494.
[4]	石敏, 禚心如, 孙碧莲, 韩国庆, 朱登明. 适用于不同款式的无监督服装动画预测[J]. 图学学报, 2024, 45(3): 539-547.
[5]	何柳, 安然, 刘姝妍, 李润岐, 陶剑, 曾照洋. 基于知识图谱的航空多模态数据组织与知识发现技术研究[J]. 图学学报, 2024, 45(2): 300-307.
[6]	宋焕生, 文雅, 孙士杰, 宋翔宇, 张朝阳, 李旭. 基于改进教师学生网络的隧道火灾检测[J]. 图学学报, 2023, 44(5): 978-987.
[7]	王佳婧, 王晨, 朱媛媛, 王笑梅. 基于民国纸币的图元素匹配检索[J]. 图学学报, 2023, 44(3): 492-501.
[8]	冯尊登, 王洪元, 林龙, 孙博言, 陈海琴. 基于多分支注意网络与相似度学习策略的无监督行人重识别[J]. 图学学报, 2023, 44(2): 280-290.
[9]	闫善武, 肖洪兵, 王瑜, 孙梅. 融合行人时空信息的视频异常检测[J]. 图学学报, 2023, 44(1): 95-103.
[10]	范震, 刘晓静, 李小波, 崔亚超. 一种对光照和遮挡鲁棒的单应性估计方法[J]. 图学学报, 2023, 44(1): 166-176.
[11]	彭国琴, 张浩, 徐丹. 基于域自适应的云南重彩画无监督情感识别[J]. 图学学报, 2022, 43(4): 641-650.
[12]	白静 , 拖继文 , 白少进 , 杨瞻源 . 基于自适应多类中心和半异构网络的三维模型草图检索 [J]. 图学学报, 2022, 43(1): 36-43.
[13]	唐静, 彭伟龙, 唐可可, 方美娥. 基于多视图网络三维形状检索的通用扰动攻击[J]. 图学学报, 2022, 43(1): 93-100.
[14]	冷珏琳, 张哲, 刘田田, 郑澎. 基于几何矩的 CAD 模型形状匹配算法及应用[J]. 图学学报, 2021, 42(4): 608-614.
[15]	王万齐 1，马宝睿 2，李倩 2，卢文龙 1，刘玉身 2 . 基于属性相似性度量的 BIM 构件聚类[J]. 图学学报, 2020, 41(2): 304-312.