自動光學檢測 (AOI) 面臨照明不均勻、陰影遮擋和相機視角變化等挑戰，導致漏報。 雖然常見的解決方案採用需要大量訓練數據的卷積神經網路 (CNN)，但這對於 AOI 中的缺陷產品樣本來說通常不切實際，從而導致數據集不平衡。 此外，複雜的人工智慧（AI）模型的可解釋性通常有限，使得用戶很難相信他們的決策。 本文介紹了一種基於可解釋的生成式分散式人工智慧的機器人系統。 提出了一種新穎的方法，將 Wasserstein 生成對抗網路 (WGAN) 與局部可解釋的模型無關解釋 (LIME) 相結合，以增強模型的可解釋性。 通過這種創新的生成對抗方法獲得的最終訓練分類模型在分散式人工智慧系統中進行了測試。 使用一系列指標來評估三個子系統的性能，包括準確度、精確度、召回率、F1 分數、接受者操作特徵曲線 (ROC)、曲線下面積和特異性。 為了驗證人工數據的真實性和多樣性，使用了 Fréchet 起始距離 (FID)、最大平均差異 (MMD)、結構相似性指數度量 (SSIM) 和峰值信噪比 (PSNR) 等指標。 實驗結果表明，源自真實樣本的訓練數據需求顯著減少了 80%，凸顯了生成的訓練數據的有效性和效率。 分散式AI從訓練時間、人工智慧算力消耗等方面進行評估。
關鍵字：自動光學檢測、可解釋人工智慧、生成對抗網路、智慧製造

Automated Optical Inspection (AOI) confronts challenges such as uneven illumination, shadow occlusion, and variations in camera viewpoint, leading to false negatives (missed reporting). Whereas common solutions employ Convolutional Neural Network (CNN) that require large amounts of training data, this is often impractical for defective product samples in AOI, resulting in imbalanced datasets. Furthermore, complex Artificial Intelligence (AI) models often have limited interpretability, making it difficult for users to trust their decisions. This paper introduces a robot system grounded in explainable generative and distributed AI. A novel approach is proposed that integrates Wasserstein Generative Adversarial Network (WGAN) with Local Interpretable Model-Agnostic Explanations (LIME) to augment the explainability of the model. The final trained classification model, obtained through this innovative generative adversarial approach, is tested within the distributed AI system. The performance of the three subsystems is evaluated using a range of metrics, including accuracy, precision, recall, F1 score, Receiver Operating Characteristic curve (ROC), area under the curve, and specificity. To address the realism and diversity of the artificial data, metrics such as Fréchet Inception Distance (FID), Maximum Mean Discrepancy (MMD), Structural Similarity Index Metric (SSIM), and Peak Signal-to-Noise Ratio (PSNR) are utilized. Experimental results demonstrate an impressive 80% reduction in training data requirements derived from real samples, highlighting the effectiveness and efficiency of the generated training data. Distributed AI is evaluated in terms of training time, and AI computing power consumption.
Keywords: Automated optical inspection, explainable artificial intelligence, generative adversarial network, smart manufacturing.

[1] R. X. Gao, X. Xu, and I. Ragai, "Advances and innovations in manufacturing systems research 2023," ed: Elsevier, 2023.
[2] A. Singh, E. F. Garcia, Y. Jeong, and M. Wiktorsson, "A Rubric for Implementing Explainable AI in Production Logistics," in Advances in Production Management Systems. Smart Manufacturing and Logistics Systems: Turning Ideas into Action: IFIP WG 5.7 International Conference, APMS 2022, Gyeongju, South Korea, September 25–29, 2022, Proceedings, Part I, 2022: Springer, pp. 190-197.
[3] M. Abd Al Rahman and A. Mousavi, "A review and analysis of automatic optical inspection and quality monitoring methods in electronics industry," IEEE Access, vol. 8, pp. 183192-183271, 2020.
[4] Y.-T. Li, P. Kuo, and J.-I. Guo, "Automatic industry PCB board DIP process defect detection system based on deep ensemble self-adaption method," IEEE Transactions on Components, Packaging and Manufacturing Technology, vol. 11, no. 2, pp. 312-323, 2020.
[5] A. Caliskan and G. Gurkan, "Design and Realization of an Automatic Optical Inspection System for PCB Solder Joints," in 2021 International Conference on INnovations in Intelligent SysTems and Applications (INISTA), 2021: IEEE, pp. 1-6.
[6] D. Li, S. Fu, Q. Zhang, Y. Mo, L. Liu, and C. Xu, "An improved pcb defect detector based on feature pyramid networks," in 2020 4th International Conference on Computer Science and Artificial Intelligence, 2020, pp. 233-239.
[7] Y. Ding et al., "An efficient industrial federated learning framework for AIoT: a face recognition application," arXiv preprint arXiv:2206.13398, 2022.
[8] Y.-Y. Chen, Y.-H. Lin, Y.-C. Hu, C.-H. Hsia, Y.-A. Lian, and S.-Y. Jhong, "Distributed real-time object detection based on edge-cloud collaboration for smart video surveillance applications," IEEE Access, vol. 10, pp. 93745-93759, 2022.
[9] A. García-Pérez et al., "CNN-based in situ tool wear detection: A study on model training and data augmentation in turning inserts," Journal of Manufacturing Systems, vol. 68, pp. 85-98, 2023.
[10] Y. Qin, X. Liu, C. Yue, M. Zhao, X. Wei, and L. Wang, "Tool wear identification and prediction method based on stack sparse self-coding network," Journal of Manufacturing Systems, vol. 68, pp. 72-84, 2023.
[11] N. Shvai, A. Hasnat, and A. Nakib, "Multiple auxiliary classifiers GAN for controllable image generation: Application to license plate recognition," IET Intelligent Transport Systems, vol. 17, no. 1, pp. 243-254, 2023.
[12] L. H. d. S. Silva, O. d. A. George, B. J. Fernandes, B. L. Bezerra, E. B. Lima, and S. C. Oliveira, "Automatic optical inspection for defective PCB detection using transfer learning," in 2019 IEEE Latin American Conference on Computational Intelligence (LA-CCI), 2019: IEEE, pp. 1-6.
[13] J. A. B. Susa, E. Mariquina, M. L. Tria, C. M. Adolfo, and J. C. Castro, "Cap-Eye-citor: A Machine Vision Inference Approach of Capacitor Detection for PCB Automatic Optical Inspection," in 2020 IEEE 7th International Conference on Engineering Technologies and Applied Sciences (ICETAS), 2020: IEEE, pp. 1-5.
[14] J. M. Rožanec, B. Fortuna, and D. Mladenić, "Knowledge graph-based rich and confidentiality preserving Explainable Artificial Intelligence (XAI)," Information Fusion, vol. 81, pp. 91-102, 2022.
[15] W. Ullah et al., "Artificial Intelligence of Things-assisted two-stream neural network for anomaly detection in surveillance Big Video Data," Future Generation Computer Systems, vol. 129, pp. 286-297, 2022.
[16] R. Lei, D. Yan, H. Wu, and Y. Peng, "A Precise Convolutional Neural Network-based Classification and Pose Prediction Method for PCB Component Quality Control," in 2022 14th International Conference on Electronics, Computers and Artificial Intelligence (ECAI), 2022: IEEE, pp. 1-6.
[17] Y.-S. Deng, A.-C. Luo, and M.-J. Dai, "Building an automatic defect verification system using deep neural network for pcb defect classification," in 2018 4th International Conference on Frontiers of Signal Processing (ICFSP), 2018: IEEE, pp. 145-149.
[18] H. Wu, R. Lei, and Y. Peng, "PCBNet: A Lightweight Convolutional Neural Network for Defect Inspection in Surface Mount Technology," IEEE Transactions on Instrumentation and Measurement, vol. 71, pp. 1-14, 2022.
[19] J. M. Rožanec et al., "Human-centric artificial intelligence architecture for industry 5.0 applications," International Journal of Production Research, pp. 1-26, 2022.
[20] Y. S. Chuo et al., "Artificial intelligence enabled smart machining and machine tools," Journal of Mechanical Science and Technology, vol. 36, no. 1, pp. 1-23, 2022.
[21] K. Zhang and H. Shen, "An Effective Multi-Scale Feature Network for Detecting Connector Solder Joint Defects," Machines, vol. 10, no. 2, p. 94, 2022.
[22] D. Li, J. Li, Y. Fan, G. Lu, J. Ge, and X. Liu, "Printed label defect detection using twice gradient matching based on improved cosine similarity measure," Expert Systems with Applications, vol. 204, p. 117372, 2022.
[23] Z. Sheng and G. Wang, "Fast Method of Detecting Packaging Bottle Defects Based on ECA-EfficientDet," Journal of Sensors, vol. 2022, 2022.
[24] Y. Miao, Z. Liu, X. Wu, and J. Gao, "Cost-sensitive Siamese network for PCB defect classification," Computational Intelligence and Neuroscience, vol. 2021, 2021.
[25] H. Zhang, L. Jiang, and C. Li, "CS-ResNet: Cost-sensitive residual convolutional neural network for PCB cosmetic defect detection," Expert Systems with Applications, vol. 185, p. 115673, 2021.
[26] M. A. N. Oz, M. Mercimek, and O. T. Kaymakci, "Anomaly localization in regular textures based on deep convolutional generative adversarial networks," Applied Intelligence, vol. 52, no. 2, pp. 1556-1565, 2022.
[27] Z. Du, L. Gao, and X. Li, "A New Contrastive GAN with Data Augmentation for Surface Defect Recognition Under Limited Data," IEEE Transactions on Instrumentation and Measurement, 2022.
[28] A. B. Haque, A. N. Islam, and P. Mikalef, "Explainable Artificial Intelligence (XAI) from a user perspective: A synthesis of prior literature and problematizing avenues for future research," Technological Forecasting and Social Change, vol. 186, p. 122120, 2023.
[29] G. Rong, Y. Xu, X. Tong, and H. Fan, "An edge-cloud collaborative computing platform for building AIoT applications efficiently," Journal of Cloud Computing, vol. 10, pp. 1-14, 2021.
[30] M. A. Guillén et al., "Performance evaluation of edge-computing platforms for the prediction of low temperatures in agriculture using deep learning," The Journal of Supercomputing, vol. 77, pp. 818-840, 2021.
[31] M. Arjovsky, S. Chintala, and L. Bottou, "Wasserstein generative adversarial networks," in International conference on machine learning, 2017: PMLR, pp. 214-223.
[32] A. K. Bedaka, S.-C. Lee, A. M. Mahmoud, Y.-S. Cheng, and C.-Y. Lin, "A camera-based position correction system for autonomous production line inspection," Sensors, vol. 21, no. 12, p. 4071, 2021.
[33] C.-H. Lin, S.-H. Wang, and C.-J. Lin, "Using convolutional neural networks for character verification on integrated circuit components of printed circuit boards," Applied Intelligence, vol. 49, no. 11, pp. 4022-4032, 2019.
[34] J. M. Runji and C.-Y. Lin, "Markerless cooperative augmented reality-based smart manufacturing double-check system: Case of safe PCBA inspection following automatic optical inspection," Robotics and Computer-Integrated Manufacturing, vol. 64, p. 101957, 2020.
[35] L. Wang, Y. Zhang, X. Xie, and X. Liu, "Application of PA-LSTM-FCN model in fault diagnosis of circuit board," in International Conference on Mechanisms and Robotics (ICMAR 2022), 2022, vol. 12331: SPIE, pp. 865-875.
[36] H. Bani-Salameh, M. Sallam, and B. Al shboul, "A deep-learning-based bug priority prediction using RNN-LSTM neural networks," e-Informatica Software Engineering Journal, vol. 15, no. 1, 2021.
[37] P. Lu, J. Jing, and Y. Huang, "MRD-net: An effective CNN-based segmentation network for surface defect detection," IEEE Transactions on Instrumentation and Measurement, vol. 71, pp. 1-12, 2022.
[38] I. Goodfellow et al., "Generative adversarial networks," Communications of the ACM, vol. 63, no. 11, pp. 139-144, 2020.
[39] Y. Qie, M. Balaghi, and N. Anwer, "Wasserstein generative adversarial networks for form defects modeling," Procedia CIRP, vol. 114, pp. 7-12, 2022.
[40] Y.-T. K. Lai and J.-S. Hu, "A texture generation approach for detection of novel surface defects," in 2018 IEEE International Conference on Systems, Man, and Cybernetics (SMC), 2018: IEEE, pp. 4357-4362.
[41] R. Singh, R. Garg, N. S. Patel, and M. W. Braun, "Generative adversarial networks for synthetic defect generation in assembly and test manufacturing," in 2020 31st Annual SEMI Advanced Semiconductor Manufacturing Conference (ASMC), 2020: IEEE, pp. 1-5.
[42] E. Thirumagal and K. Saruladha, "Lung cancer classification using exponential mean saturation linear unit activation function in various generative adversarial network models," International Journal of Imaging Systems and Technology, vol. 32, no. 4, pp. 1414-1428, 2022.
[43] O. Nwafor, E. Okafor, A. Aboushady, C. Nwafor, and C. Zhou, "Explainable Artificial Intelligence for Prediction of Non-Technical Losses in Electricity Distribution Networks," IEEE Access, 2023.
[44] M. T. Ribeiro, S. Singh, and C. Guestrin, "" Why should i trust you?" Explaining the predictions of any classifier," in Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining, 2016, pp. 1135-1144.
[45] F. Doshi-Velez and B. Kim, "Towards a rigorous science of interpretable machine learning," arXiv preprint arXiv:1702.08608, 2017.
[46] Q.-s. Zhang and S.-C. Zhu, "Visual interpretability for deep learning: a survey," Frontiers of Information Technology & Electronic Engineering, vol. 19, no. 1, pp. 27-39, 2018.
[47] S. R. Hong, J. Hullman, and E. Bertini, "Human factors in model interpretability: Industry practices, challenges, and needs," Proceedings of the ACM on Human-Computer Interaction, vol. 4, no. CSCW1, pp. 1-26, 2020.
[48] X. Wei, "Joint Resource Management and Task Scheduling for Mobile Edge Computing," Temple University, 2023.
[49] B. Ren, Y. Chen, and F. Wang, "Application massive data processing platform for smart manufacturing based on optimization of data storage," ACM Transactions on Management Information Systems, vol. 13, no. 4, pp. 1-20, 2023.
[50] X. Hu, Y. Li, L. Jia, and M. Qiu, "A novel two-stage unsupervised fault recognition framework combining feature extraction and fuzzy clustering for collaborative AIoT," IEEE Transactions on Industrial Informatics, vol. 18, no. 2, pp. 1291-1300, 2021.
[51] L. Chuanhong, G. Caili, Y. Yang, F. Chunyan, S. Qizheng, and C. Jiujiu, "Intelligent task-oriented semantic communication method in artificial intelligence of things," Journal on Communication/Tongxin Xuebao, vol. 42, no. 11, 2021.
[52] D.-M. Tsai, Y.-Q. Huang, and W.-Y. Chiu, "Deep learning from imbalanced data for automatic defect detection in multicrystalline solar wafer images," Measurement Science and Technology, vol. 32, no. 12, p. 124003, 2021.
[53] T. Cheng, X. Liu, L. Qin, M. Lu, C. Xiao, and S. Li, "A Practical Micro Fringe Projection Profilometry for 3-D Automated Optical Inspection," IEEE Transactions on Instrumentation and Measurement, vol. 71, pp. 1-13, 2022.
[54] T.-W. Tang, W.-H. Kuo, J.-H. Lan, C.-F. Ding, H. Hsu, and H.-T. Young, "Anomaly detection neural network with dual auto-encoders GAN and its industrial inspection applications," Sensors, vol. 20, no. 12, p. 3336, 2020.
[55] Y. Chen, T. Zhao, P. Cheng, M. Ding, and C. W. Chen, "Joint front–edge–cloud iovt analytics: Resource-effective design and scheduling," IEEE Internet of Things Journal, vol. 9, no. 23, pp. 23941-23953, 2022.
[56] M.-F. R. Lee, C. W. de Silva, E. A. Croft, and Q. J. Wu, "Machine vision system for curved surface inspection," Machine vision and Applications, vol. 12, pp. 177-188, 2000.
[57] H. Chen, L. Hou, G. K. Zhang, and S. Wu, "Using Context-Guided data Augmentation, lightweight CNN, and proximity detection techniques to improve site safety monitoring under occlusion conditions," Safety science, vol. 158, p. 105958, 2023.
[58] J. Arrieta, C. Mera, and A. Espinosa, "Evaluation of Weakly Supervised Learning Paradigms on Automatic Visual Inspection," IEEE Latin America Transactions, vol. 18, no. 06, pp. 1017-1025, 2020.
[59] Z. Zheng, S. Zhang, J. Shen, Y. Shao, and Y. Zhang, "A two-stage CNN for automated tire defect inspection in radiographic image," Measurement Science and Technology, vol. 32, no. 11, p. 115403, 2021.
[60] W. Yang, Y. Xiao, H. Shen, and Z. Wang, "An effective data enhancement method of deep learning for small weld data defect identification," Measurement, vol. 206, p. 112245, 2023.
[61] J. Schubeck, D. Koblah, U. J. Botero, and D. Forte, "A comprehensive taxonomy of PCB defects," ed.
[62] Q. Ling and N. A. M. Isa, "Printed Circuit Board Defect Detection Methods Based on Image Processing, Machine Learning and Deep Learning: A Survey," IEEE Access, 2023.
[63] V. U. Sankar, G. Lakshmi, and Y. S. Sankar, "A review of various defects in PCB," Journal of Electronic Testing, vol. 38, no. 5, pp. 481-491, 2022.
[64] G. Besnard, B.-Y. Nguyen, and C. Maleville, "Smart Cut™ technology: from Substrate Enginnering to Advanced 3D Integration," in 2022 International Conference on IC Design and Technology (ICICDT), 2022: IEEE, pp. 81-83.
[65] A. Ibrahim, S. Tominaga, and T. Horiuchi, "Spectral imaging method for material classification and inspection of printed circuit boards," Optical Engineering, vol. 49, no. 5, pp. 057201-057201-10, 2010.
[66] Q. Xu et al., "An empirical study on evaluation metrics of generative adversarial networks," arXiv preprint arXiv:1806.07755, 2018.
[67] M. Ikuta and J. Zhang, "TextureWGAN: texture preserving WGAN with multitask regularizer for computed tomography inverse problems," Journal of Medical Imaging, vol. 10, no. 2, pp. 024003-024003, 2023.