PT-MFR： a CAD model machining feature recognition method based on Point Transformer

doi:10.7523/j.ucas.2024.041

Abstract

Abstract:

Machining feature recognition is crucial in computer-aided design （CAD） and manufacturing （CAM） as it serves as a vital link between CAD and CAM systems. Researchers have proposed two types of machining feature recognition methods： rule-based and learning-based， with the latter showing superior performance and garnering more attention. However， existing recognition methods face challenges such as insufficient utilization of geometric information， inaccurate machining feature localization， and complexity in the instance segmentation process. To address these issues， this paper proposes PT-MFR， a CAD model machining feature recognition method based on Point Transformer. It performs two tasks： semantic segmentation and instance segmentation， predicting the semantic category of machining features for each face and calculating face similarity to segment the machining feature instances. The results of both tasks are integrated to obtain the machining feature recognition results. Experimental results demonstrate that the proposed method outperforms other methods.

Key words: machining feature recognition, point cloud, neural network, Point Transformer

CLC Number:

TP391

Haochen HE, Zheng FANG, Zhengda LU, Jun XIAO, Ying WANG. PT-MFR： a CAD model machining feature recognition method based on Point Transformer[J]. Journal of University of Chinese Academy of Sciences, 2026, 43(1): 115-124.

Figures/Tables 11

Fig.1 Examples of machining features in CAD models

Fig.2 Examples of machining feature recognition error

Fig.3 Architecture of the proposed network

Fig.4 Some generated CAD models

Table 1 Results of the comparative experiment

网络	语义分割			实例分割
网络	$A c c$	$m A c c$	mIoU	Acc	F1	mIoU
Cao 等^[29]提出的网络	53.44	33.34	43.90	—	—	—
Hierarchical CADNet^[8]	83.53	83.63	72.65	—	—	—
ASIN^[4]	78.13	76.44	65.09	10.88	6.81	5.77
AAGNet^[13]	88.40	89.07	79.99	53.20	60.52	55.56
PT-MFR	90.36	89.29	82.81	60.56	70.25	63.24

Table 1 Results of the comparative experiment

网络	语义分割			实例分割
网络	$A c c$	$m A c c$	mIoU	Acc	F1	mIoU
Cao 等^[29]提出的网络	53.44	33.34	43.90	—	—	—
Hierarchical CADNet^[8]	83.53	83.63	72.65	—	—	—
ASIN^[4]	78.13	76.44	65.09	10.88	6.81	5.77
AAGNet^[13]	88.40	89.07	79.99	53.20	60.52	55.56
PT-MFR	90.36	89.29	82.81	60.56	70.25	63.24

Table 2 Results of the feature extraction module ablation experiment

特征提取模块	语义分割			实例分割
特征提取模块	$A c c$	$m A c c$	mIoU	Acc	F1	mIoU
PointNet^[35]	72.10	75.53	60.97	40.54	53.95	47.14
DGCNN^[38]	80.69	82.45	71.95	34.46	50.22	44.08
Sparse Conv^[39]	74.97	73.59	61.91	41.70	54.10	47.93
PT-MFR	90.36	89.29	82.81	60.56	70.25	63.24

Table 2 Results of the feature extraction module ablation experiment

特征提取模块	语义分割			实例分割
特征提取模块	$A c c$	$m A c c$	mIoU	Acc	F1	mIoU
PointNet^[35]	72.10	75.53	60.97	40.54	53.95	47.14
DGCNN^[38]	80.69	82.45	71.95	34.46	50.22	44.08
Sparse Conv^[39]	74.97	73.59	61.91	41.70	54.10	47.93
PT-MFR	90.36	89.29	82.81	60.56	70.25	63.24

Table 3 Results of the face position ablation experiment

面位置	语义分割			实例分割
面位置	$A c c$	$m A c c$	mIoU	Acc	F1	mIoU
删除面位置	89.50	87.82	81.57	37.54	51.26	45.43
PT-MFR	90.36	89.29	82.81	60.56	70.25	63.24

Table 3 Results of the face position ablation experiment

面位置	语义分割			实例分割
面位置	$A c c$	$m A c c$	mIoU	Acc	F1	mIoU
删除面位置	89.50	87.82	81.57	37.54	51.26	45.43
PT-MFR	90.36	89.29	82.81	60.56	70.25	63.24

Table 4 Results of the loss function ablation experiment

损失函数	语义分割			实例分割
损失函数	$A c c$	$m A c c$	mIoU	Acc	F1	mIoU
MSE损失	85.12	87.71	79.35	10.07	6.69	5.33
PT-MFR	90.36	89.29	82.81	60.56	70.25	63.24

Table 4 Results of the loss function ablation experiment

损失函数	语义分割			实例分割
损失函数	$A c c$	$m A c c$	mIoU	Acc	F1	mIoU
MSE损失	85.12	87.71	79.35	10.07	6.69	5.33
PT-MFR	90.36	89.29	82.81	60.56	70.25	63.24

Table 5 Results of loss function weight ablationexperiment

$L 1$ ∶ $L 2$	语义分割			实例分割
$L 1$ ∶ $L 2$	$A c c$	$m A c c$	mIoU	Acc	F1	mIoU
1∶1	90.26	89.63	82.76	59.21	69.33	62.21
1∶5	90.94	89.87	83.68	59.91	69.11	62.16
1∶10	90.36	89.29	82.81	60.56	70.25	63.24
1∶20	86.49	81.85	77.20	61.83	71.06	64.04
1∶50	86.92	82.97	78.02	60.78	70.13	63.16

Table 5 Results of loss function weight ablationexperiment

$L 1$ ∶ $L 2$	语义分割			实例分割
$L 1$ ∶ $L 2$	$A c c$	$m A c c$	mIoU	Acc	F1	mIoU
1∶1	90.26	89.63	82.76	59.21	69.33	62.21
1∶5	90.94	89.87	83.68	59.91	69.11	62.16
1∶10	90.36	89.29	82.81	60.56	70.25	63.24
1∶20	86.49	81.85	77.20	61.83	71.06	64.04
1∶50	86.92	82.97	78.02	60.78	70.13	63.16

Fig.5 Comparative case study on model A

Fig.6 Results of machining feature recognition in real-world model

References 39

[1]	边培莹. 机械CAD/CAM原理及应用［M］. 北京：机械工业出版社， 2020.
[2]	王秋明，高慧颖，刘科成. 基于过程方法的现代集成制造模式下质量管理系统模型［J］. 中国科学院研究生院学报， 2012， 29（5）： 686-692. DOI： 10.7523/j.issn.2095-6134.2012.5.016 .
[3]	Shi Y， Zhang Y C， Xia K S， et al. A critical review of feature recognition techniques［J］. Computer-Aided Design and Applications， 2020， 17（5）： 861-899. DOI： 10.14733/cadaps.2020.861-899 .
[4]	Zhang H， Zhang S S， Zhang Y J， et al. Machining feature recognition based on a novel multi-task deep learning network［J］. Robotics and Computer-Integrated Manufacturing， 2022， 77： 102369. DOI： 10.1016/j.rcim.2022.102369 .
[5]	王泽隆，徐向辉，张雷. 基于仿真SAR图像深度迁移学习的自动目标识别［J］. 中国科学院大学学报， 2020， 37（4）： 516-524. DOI： 10.7523/j.issn.2095-6134.2020.04.011 .
[6]	孟曦婷，计璐艳，赵永超，等. 基于深度学习的多尺度导弹发射井目标检测［J］. 中国科学院大学学报， 2021， 38（6）： 800-808. DOI： 10.7523/j.issn.2095-6134.2021.06.010 .
[7]	Zhang Y， Zhang Y S， He K W， et al. Intelligent feature recognition for STEP-NC-compliant manufacturing based on artificial bee colony algorithm and back propagation neural network［J］. Journal of Manufacturing Systems， 2022， 62： 792-799. DOI： 10.1016/j.jmsy.2021.01.018 .
[8]	Colligan A R， Robinson T T， Nolan D C， et al. Hierarchical CADNet： learning from B-reps for machining feature recognition［J］. Computer-Aided Design， 2022， 147： 103226. DOI： 10.1016/j.cad.2022.103226 .
[9]	Shi P Z， Qi Q F， Qin Y C， et al. Intersecting machining feature localization and recognition via single shot multibox detector［J］. IEEE Transactions on Industrial Informatics， 2021， 17（5）： 3292-3302. DOI： 10.1109/TII.2020.3030620 .
[10]	Shi P Z， Qi Q F， Qin Y C， et al. Highly interacting machining feature recognition via small sample learning［J］. Robotics and Computer-Integrated Manufacturing， 2022， 73： 102260. DOI： 10.1016/j.rcim.2021 .
[11]	Zhang Z B， Jaiswal P， Rai R. FeatureNet： machining feature recognition based on 3D convolution neural network［J］. Computer-Aided Design， 2018， 101： 12-22. DOI： 10.1016/j.cad.2018.03.006 .
[12]	Shi Y， Zhang Y C， Harik R. Manufacturing feature recognition with a 2D convolutional neural network［J］. CIRP Journal of Manufacturing Science and Technology， 2020， 30： 36-57. DOI： 10.1016/j.cirpj.2020.04.001 .
[13]	Wu H J， Lei R S， Peng Y B， et al. AAGNet： a graph neural network towards multi-task machining feature recognition［J］. Robotics and Computer-Integrated Manufacturing， 2024， 86： 102661. DOI： 10.1016/j.rcim.2023.102661 .
[14]	Wu X Y， Lao Y X， Jiang L， et al. Point transformer V2： grouped vector attention and partition-based pooling［EB/OL］. arXiv 2022： 2210.05666.（2022-10-11）［2024-04-28］. .
[15]	Zhao H S， Jiang L， Jia J Y， et al. Point transformer［C］//2021 IEEE/CVF International Conference on Computer Vision （ICCV）. Montreal， QC， Canada. IEEE， 2021： 16239-16248. DOI： 10.1109/ICCV48922.2021.01595 .
[16]	Wu X Y， Jiang L， Wang P S， et al. Point transformer V3： simpler， faster， stronger［EB/OL］. arXiv 2023： 2312.10035. （2023-12-15）［2024-04-28］. .
[17]	Ji Q， Marefat M M. Bayesian approach for extracting and identifying features［J］. Computer-Aided Design， 1995， 27（6）： 435-454. DOI： 10.1016/0010-4485（95）00017-L .
[18]	Wang Q M， Yu X H. Ontology based automatic feature recognition framework［J］. Computers in Industry， 2014， 65（7）： 1041-1052. DOI： 10.1016/j.compind.2014.04.004 .
[19]	Zhang Y Z， Luo X F， Zhang B Y， et al. Semantic approach to the automatic recognition of machining features［J］. The International Journal of Advanced Manufacturing Technology， 2017， 89（1）： 417-437. DOI： 10.1007/s00170-016-9056-8 .
[20]	Mandorli F， Borgo S， Wiejak P. From form features to semantic features in existing MCAD： an ontological approach［J］. Advanced Engineering Informatics， 2020， 44： 101088. DOI： 10.1016/j.aei.2020.101088 .
[21]	Trika S N， Kashyap R L. Geometric reasoning for extraction of manufacturing features in iso-oriented polyhedrons［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 1994， 16（11）： 1087-1100. DOI： 10.1109/34.334388 .
[22]	Li Y G， Ding Y F， Mou W P， et al. Feature recognition technology for aircraft structural parts based on a holistic attribute adjacency graph［J］. Proceedings of the Institution of Mechanical Engineers， Part B： Journal of Engineering Manufacture， 2010， 224（2）： 271-278. DOI： 10.1243/09544054jem1634 .
[23]	刘云华，陈立平，钟毅芳，等. 基于痕迹的特征模型解释技术研究［J］. 计算机学报， 2000， 23（9）： 960-965. DOI： 10.3321/j.issn：0254-4164.2000.09.013 .
[24]	Li Z， Lin J， Zhou D C， et al. Local symmetry based hint extraction of B-rep model for machining feature recognition［C］//ASME 2018 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference， August 26–29， 2018， Quebec City， Quebec， Canada. 2018 DOI： 10.1115/DETC2018-85509 .
[25]	Kim B C， Mun D. Stepwise volume decomposition for the modification of B-rep models［J］. The International Journal of Advanced Manufacturing Technology， 2014， 75（9）： 1393-1403. DOI： 10.1007/s00170-014-6210-z .
[26]	Kataraki P S， Abu Mansor M S. Auto-recognition and generation of material removal volume for regular form surface and its volumetric features using volume decomposition method［J］. The International Journal of Advanced Manufacturing Technology， 2017， 90（5）： 1479-1506. DOI： 10.1007/s00170-016-9394-6 .
[27]	Rameshbabu V， Shunmugam M S. Hybrid feature recognition method for setup planning from STEP AP-203［J］. Robotics and Computer-Integrated Manufacturing， 2009， 25（2）： 393-408. DOI： 10.1016/j.rcim.2007.09.014 .
[28]	Guo L， Zhou M， Lu Y Q， et al. A hybrid 3D feature recognition method based on rule and graph［J］. International Journal of Computer Integrated Manufacturing， 2021， 34（3）： 257-281. DOI： 10.1080/0951192x.2020.1858507 .
[29]	Cao W J， Robinson T， Hua Y， et al. Graph representation of 3D CAD models for machining feature recognition with deep learning［C］//ASME 2020 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference， August 17–19， 2020， Virtual， Online. 2020 DOI： 10.1115/DETC2020-22355 .
[30]	Jian C F， Li M， Qiu K Y， et al. An improved NBA-based STEP design intention feature recognition［J］. Future Generation Computer Systems， 2018， 88： 357-362. DOI： 10.1016/j.future.2018.05.033 .
[31]	Yeo C， Kim B C， Cheon S， et al. Machining feature recognition based on deep neural networks to support tight integration with 3D CAD systems［J］. Scientific Reports， 2021， 11（1）： 22147. DOI： 10.1038/s41598-021-01313-3 .
[32]	石叶楠，郑国磊. 三种用于加工特征识别的神经网络方法综述［J］. 航空学报， 2019， 40（9）： 022840. DOI： 10.7527/S1000-6893.2019.22840 .
[33]	Yao X H， Wang D， Yu T， et al. A machining feature recognition approach based on hierarchical neural network for multi-feature point cloud models［J］. Journal of Intelligent Manufacturing， 2023， 34（6）： 2599-2610. DOI： 10.1007/s10845-022-01939-8 .
[34]	Qi C R， Yi L， Su H， et al. PointNet++： deep hierarchical feature learning on point sets in a metric space［EB/OL］. arXiv 2017： 1706.02413.（2017-06-07）［2024-04-28］. .
[35]	Charles R Q， Hao S， Mo K C， et al. PointNet： deep learning on point sets for 3D classification and segmentation［C］//2017 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）. Honolulu， HI， USA. IEEE， 2017： 77-85. DOI： 10.1109/CVPR.2017.16 .
[36]	Vaswani A， Shazeer N， Parmar N， et al. Attention is all you need［EB/OL］. arXiv 2017： 1706.03762. （2017-06-12）［2024-04-28］. .
[37]	Kingma D P， Ba J. Adam： a method for stochastic optimization［EB/OL］. arXiv 2014： 1412.6980.（2014-12-22）［2024-04-28］. .
[38]	Wang Y， Sun Y B， Liu Z W， et al. Dynamic graph CNN for learning on point clouds［J］. ACM Transactions on Graphics， 2019， 38（5）： 146. DOI： 10.1145/3326362 .
[39]	Graham B， Engelcke M， van der Maaten L. 3D semantic segmentation with submanifold sparse convolutional networks［C］//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City， UT， USA. IEEE， 2018： 9224-9232. DOI： 10.1109/CVPR.2018.00961 .