電腦輔助製程規劃是通過自動化規劃零件加工製造的過程，減少手動干預，加快生產計劃的生成速度，即因應零件的改動，快速調整新的製程需求，降低生產成本。電腦輔助製程規劃的其中一個重要部為特徵辨識，其目的為將3D幾何模型的拓樸資訊轉換為加工特徵資訊，減少加工人員的作業時間與精力。雖然目前已有多種特徵辨識方法被提出，但沒有一種是針對具有高度複雜性的單一特徵，這導致在這些方法或神經網路模型難以應用在實際特徵當中，因此本研究提出建立高度複雜性之單一特徵的3D幾何模型的方法，並以圖基法進行特徵提取，將特些特徵提取後的面以特定的編碼規則進行神經網路的訓練，接著使用訓練後的神經網路進行特徵類別的預測。
本文詳述了自動化建立高度複雜性之單一特徵的3D幾何模型的方法、自動化特徵搜尋的方法、特徵編碼之規則及一個有效運用這些特徵代碼的神經網路之架構，並以幾個實例驗證之結果進行探討。

Computer-aided process planning is a process that automates the planning of part manufacturing, reducing manual intervention, accelerating the generation speed of production plans, and swiftly adjusting to new process requirements in response to changes in parts, thereby lowering production costs. One crucial aspect of computer-aided process planning is feature recognition, which aims to convert the topological information of 3D geometric models into machining feature information, thereby reducing the operational time and effort of machining personnel. Although various feature recognition methods have been proposed, none specifically target highly complex individual features. This limitation makes it challenging to apply these methods or neural network models to practical features. Therefore, this research proposes a method for constructing 3D geometric models of highly complex individual features and employs graph-based techniques for feature extraction. The extracted features are then trained using a neural network with specific encoding rules. Subsequently, the trained neural network is used to predict feature categories.
This thesis details the automated method for constructing 3D geometric models of highly complex individual features, an automated feature search method, rules for feature encoding, and an effective neural network architecture that utilizes these feature codes. The results of several case studies are discussed for validation.

[1] Joshi Sanjay and Tien-Chien Chang (1988), “Graph-based heuristics for recognition of machined features from a 3D solid model,” Computer -Aided Design, Vol. 20, No. 2, pp. 58-66.
[2] Laakko Timo and Martti Mäntylä (1993), “Feature modelling by incremental feature recognition,” Computer-Aided Design, Vol. 25, No. 8, pp. 479-492.
[3] C. F. Yuen and Venuvinod Patrik (1999), “Geometric feature recognition: coping with the complexity and infinite variety of features,” International Journal of Computer Integrated Manufacturing, Vol. 12 No. 5, pp. 439-452.
[4] Shu-Ming Gao and Jami J. Shah (1998), “Automatic recognition of interacting machining features based on minimal condition subgraph,” Computer-Aided Design, Vol.30, No. 9, pp.727-739.
[5] Ying-Guang Li, Y. F. Ding, H Guo and Wen-Ping Mou (2010), “Feature recognition technology for aircraft structural parts based on a holistic attribute adjacency graph,” Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture, Vol. 224, No. 2, pp. 271-278.
[6] Na Cai, Soumiya Bendjebla, Sylvain Lavernhe, Charyar Mehdi-Souzani, Nabil Anwer (2018), “Freeform machining feature recognition with manufacturability analysis,” Procedia CIRP, Vol. 72, pp. 1475-1480.
[7] Iain A. Donaldson and Jonathan R. Corney (1993), “Rule-based feature recognition for 2· 5D machined components,” International Journal of Computer Integrated Manufacturing, Vol. 6, No. 1-2, pp. 51-64.
[8] Hai-yan Li, Yun-Bao Huang, Yu-Hang Sun and Li-Ping Chen (2015), “Hint-based generic shape feature recognition from three-dimensional B-rep models,” Advances in Mechanical Engineering, Vol. 7.4: 1687814015582082.
[9] Yong-Se Kim (1992), “Recognition of form features using convex decomposition,” Computer-Aided Design, Vol. 24, No. 9, pp. 461-476.
[10] Betkier Igor ,Oszczypała Mateusz ,Pobożniak Janusz , Sobieski Sergiusz, Betkier Przemysław (2023), “PocketFinderGNN: A manufacturing feature recognition software based on Graph Neural Networks (GNNs) using PyTorch Geometric and NetworkX,” SoftwareX, Vol. 23: 101466.
[11] Hong-Jin Wu, Ruo-Shan Lei, Yi-Bing Peng and Liang Gao (2024), “AAGNet: A graph neural network towards multi-task machining feature recognition,” Robotics and Computer-Integrated Manufacturing, Vol. 86, 102661.
[12] Prabhakar Shashikanth and Mark R. Henderson (1993), “Automatic form-feature recognition using neural-network-based techniques on boundary representations of solid models,” Computer-Aided Design, Vol. 24, No. 7, pp. 381-393.
[13] Sunil V. B. and S. S. Pande (2009), “Automatic recognition of machining features using artificial neural networks,” International Journal of Advanced Manufacturing Technology, Vol. 41, pp. 932-947.
[14] Yu Zhang, Yong-Sheng Zhang, Kai-Wen He, Dong-Sheng Li, Xun Xu and Ya-Dong Gong (2022), “Intelligent feature recognition for STEP-NC-compliant manufacturing based on artificial bee colony algorithm and back propagation neural network,” Journal of Manufacturing Systems, Vol. 62, pp. 792-799.
[15] Pei-Zhi Shi, Qun-Fen Qi, Yu-Chu Qin, Paul J. Scott and Xiang-Qian Jiang (2020), “A novel learning-based feature recognition method using multiple sectional view representation,” Journal of Intelligent Manufacturing, Vol. 31, pp. 1291-1309.
[16] Pei-Zhi Shi, Qun-Fen Qi, Yu-Chu Qin, Paul J. Scott and Xiang-Qian Jiang (2020), “Intersecting machining feature localization and recognition via single shot multibox detector,” IEEE Transactions on Industrial Informatics, Vol. 17, No. 5, pp. 3292-3302.
[17] Pei-Zhi Shi, Qun-Fen Qi, Yu-Chu Qin, Paul J. Scott and Xiang-Qian Jiang (2022),“Highly interacting machining feature recognition via small sample learning,” Robotics and Computer-Integrated Manufacturing, Vol. 73, 102260.
[18] Maturana Daniel and Sebastian Scherer. (2015), “Voxnet: A 3d convolutional neural network for real-time object recognition,” IEEE/RSJ International Conference on Intelligent Robots and Systems, September 28-October 2, Hamburg, Germany, pp. 922-928.
[19] Zhi-Bo Zhang, Prakhar Jaiswal and Rahul Rai (2018), “Featurenet: Machining feature recognition based on 3d convolution neural network,” Computer-Aided Design, Vol. 101, pp. 12-22.
[20] Xin-Hua Yao, Di Wang, Tao Yu, Cong-Cong Luan and Jian-Zhong Fu (2023), “A machining feature recognition approach based on hierarchical neural network for multi-feature point cloud models,” Journal of Intelligent Manufacturing, Vol. 34, No. 6, pp. 2599-2610.
[21] Ruo-Shan Lei, Hong-Jin Wu and Yi-Ying Peng (2020), “MFPointNet: A Point Cloud-Based Neural Network Using Selective Downsampling Layer for Machining Feature Recognition,” Machines, Vol. 10, No. 12, 1165.
[22] Hang Zhang, Shu-Sheng Zhang, Ya-Jun Zhang, Jia-Chen Liang and Zhen Wang (2022), “Machining feature recognition based on a novel multi-task deep learning network,” Robotics and Computer-Integrated Manufacturing, Vol. 77, 102369.
[23] Rana Hanocka, Amir Hertz, Noa Fish, Raja Giryes, Shachar Fleishman and Daniel Cohen-Or (2019), “Meshcnn: a network with an edge,” ACM Transactions on Graphics,Vol. 38, No. 4, pp. 1-12.
[24] Yu-Tong Feng, Yi-Fan Feng, Hao-Xuan You, Xi-Bin Zhao, Yue Gao (2019), “Meshnet: Mesh neural network for 3d shape representation,” Proceedings of the AAAI Conference on Artificial Intelligence, January 27-February 1, Honolulu, Hawaii, USA, Vol. 33, No. 01, pp. 8279-8286.
[25] Joseph G. Lambourne, Karl D.D. Willis, Pradeep Kumar Jayaraman, Aditya Sanghi, Peter Meltzer and Hooman Shayani (2021), “Brepnet: A topological message passing system for solid models,” Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 20-25, Nashville, TN, USA, pp. 12773-12782.
[26] Pradeep Kumar Jayaraman, Aditya Sanghi, Joseph G. Lambourne, Karl D.D. Willis, Thomas Davies, Hooman Shayani and Nigel Morris (2023), “Uv-net: Learning from boundary representations,” Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 20-25, Nashville, TN, USA, pp. 11703-11712.
[27] Andrew R. Colligan, Trevor T. Robinson, Declan C. Nolan, Yang Hua, Weijuan Cao (2022), “Hierarchical cadnet: Learning from b-reps for machining feature recognition,” Computer-Aided Design, Vol. 147, 03226.
[28] Peng-Yu Wang, Wen-An Yang and You-Peng You (2023), “A hybrid learning framework for manufacturing feature recognition using graph neural networks,” Journal of Manufacturing Processes, Vol.85, pp. 387-404.
[29] Wei-Juan Cao, T. Robinson, Yang Hua, F. Boussuge, Andrew R. Colligan and Wan-Bin Pan (2020), “Graph representation of 3D CAD models for machining feature recognition with deep learning,” International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, August 17-19, Virtual, Online, Vol. 84003, American Society of Mechanical Engineers.
[30] Karl D.D. Willis, Yewen Pu, Jieliang Luo, Hang Chu, Tao Du, Joseph G. Lambourne, Armando Solar-Lezama and Wojciech Matusik (2021), “Fusion 360 gallery: A dataset and environment for programmatic cad construction from human design sequences,” ACM Transactions on Graphics, Vol. 40, No. 4, pp. 1-24.
[31] Rundi Wu, Chang Xiao and Changxi Zheng (2021), “Deepcad: A deep generative network for computer-aided design models,” Proceedings of the IEEE/CVF International Conference on Computer Vision, October 10-17, Montreal, QC, Canada, pp. 6772-6782.
[32] Open CASCADE Technology, Retrieved from https://dev.opencascade.org/
[33] PythonOCC, Retrieved from https://dev.opencascade.org/project/pythonocc
[34] NX Open, Retrieved from
NX OPEN API Automation & Customization Programming (nx-open.com)
[35] PyTorch, Retrieved from
https://pytorch.org/get-started/previous-versions/