簡易檢索 / 詳目顯示

研究生: 古昊中
Hao-Chung Ku
論文名稱: 具目標導向擬人化學習之物聯網廣域監控
Goal-driven Anthropomorphic Learning for IoT-enabled Wide-area Monitoring
指導教授: 陸敬互
Ching-Hu Lu
口試委員: 廖峻鋒
Chun-Feng Liao
鍾聖倫
Sheng-Luen Chung
馬尚彬
Shang-Pin Ma
蘇順豐
Shun-Feng Su
陸敬互
Ching-Hu Lu
學位類別: 碩士
Master
系所名稱: 電資學院 - 電機工程系
Department of Electrical Engineering
論文出版年: 2020
畢業學年度: 108
語文別: 中文
論文頁數: 88
中文關鍵詞: 擬人化學習自我調整學習特徵濾除關鍵影格偵測區塊鏈智慧合約邊緣運算物聯網
外文關鍵詞: anthropomorphic learning, self-regulated learning, feature filtering, key frame detection, blockchain, smart contract, edge computing, Internet of Things
相關次數: 點閱:479下載:0
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報
  • 隨著物聯網時代的興起,大量具邊緣運算能力的攝影機 (本研究稱為邊緣攝影機)於廣域監控已逐漸廣泛提供串流影像之即時處理,並具有擬人化的自我評估與決策的應用。因此,機器學習與自動化功能結合邊緣運算之相關研究議題也更受重視。然而,邊緣攝影機在既有的擬人化廣域監控上,在輸入串流影像的隱私即時處理上未能依照不同使用者所顧慮的特徵提供客製化之影像濾除來給予隱私保護,同時也無法避免原始影像資訊遭受破壞。其次,當串流影像即時偵測產生概念漂移時,偵測模型通常未能偵測具有重要以及代表性之關鍵影像,導致既有暫存之影像代表性不足且占用儲存資源。最後,在根據偵測結果提供服務時,系統未考量多位使用者間如何協商來自動分配資源,以確保每位使用者的服務品質得以持續維持。因此,針對上述問題,本研究在自我調整學習的框架上提出對應的技術與解決方案。首先,在上述未依不同使用者客製化顧慮特徵濾除問題上,本研究提出「基於生成對抗模型之客製化影像顧慮特徵濾除」,其藉由生成對抗網路將使用者所顧慮之特徵進行訓練,並根據不同任務之需求,動態地配置對應之特徵濾除模型,以降低顧慮特徵資訊遭受其他辨識系統偵測之風險。其次,針對上述未能偵測概念飄移之關鍵影像問題,本研究提出「基於關鍵影格之模型即時調適」,針對模型偵測串流影像概念飄移之結果,藉由影像相似度保留關鍵影像,使邊緣攝影機所需之影像暫存空間得以減少,並且增加調適之效率。最後,針對上述未考量多位使用者服務品質確保問題,本研究提出「基於智慧合約之可磋商服務品質確保」,其結合區塊鏈的智慧合約提供自動磋商程序,並且根據磋商結果,於服務期間在邊緣攝影機進行自主資源管理。實驗結果顯示,「基於生成對抗模型之客製化影像顧慮特徵濾除」在本研究濾除膚色與性別特徵的情境下,生成影像在顧慮特徵的對抗成效能可以平均提升24.26%,藉此可以避免顧慮特徵遭受辨識,導致與學習目標衝突。而「基於關鍵影格之模型即時調適」可在損失小於2%辨識準確率情況下,加速52%模型整體調適時間,以及節省54.7%邊緣攝影機所需之暫存空間。最後,「基於智慧合約之可磋商服務品質確保」,本研究設計了示範應用,透過智慧合約與使用者進行服務磋商,並擔任服務提供期間的品質監控與監督角色,有效降低服務提供者的執行與維護負擔,在兼顧雙方效益下確保服務品質。


    With the rise of the Internet of Things, a large number of IoT-enabled cameras (one with the ability to leverage edge intelligence, hereafter referred to edge camera) have widely provided real-time processing of streaming images. Now, they have anthropomorphic self-evaluation and decision support in wide-area monitoring, so machine learning and system automation incorporating edge computing have attracted increasing attention. However, in previous studies on anthropomorphic wide-area monitoring, the edge cameras cannot effectively filter out the concerned features in streaming images based on different user requirements to provide privacy protection. Next, when any concept drift occurs, the detection models cannot efficiently detect important and representative images to save storage space. Finally, when providing services based on the detection results, the systems did not consider how to negotiate and maintain quality of service (QoS) for multiple users via autonomous adjustment. To address above issues, our study proposes corresponding technologies and solutions based on the self-regulated learning framework. To address these three problem, our study first proposes GAN (generative adversarial network)-based customized concerned feature filtering to screen out user’s concerned features via dynamically deployed models according to different user requirements. Next, our study proposes key-frame-oriented model adaptation by detecting the similarity among images, so the demand of image storage on an edge camera can be effectively reduced meanwhile improving adaptation efficiency. Finally, our study proposes smart-contract-based negotiable QoS insurance to incorporate smart contract on a blockchain to provide negotiation procedures, which then triggers autonomous resource management on the edge camera based on the negotiation results. The experimental results show that the customized concerned feature filtering for skin color and gender features can improve the adversarial effect of concerned features by 24.26% on average, thus avoiding the recognition results deviating from the learning goal. The key-frame-oriented model adaptation can accelerate the overall adjustment time by 52% and save the required storage space by 54.7% at the cost of only compromising precision 2% at most. Finally, a real-life scenario has been implemented to demonstrate smart-contract-based negotiable QoS insurance. The smart contract serves as an agent to negotiate with users, and it also plays the role of QoS monitoring and insurance during service provisioning, thus effectively greatly reducing human interventions for service providers and and customers.

    中文摘要 i Abstract iii 致謝 v 第一章 簡介 1 1.1 研究動機 1 1.2 文獻探討 6 1.2.1 問題一:未依使用者客製化顧慮濾除之處理方針 7 1.2.2 問題二:未能有效偵測概念飄移之關鍵影格 10 1.2.3 問題三:制式化資源分配無法確保服務品質 11 1.3 本研究貢獻與文章架構 14 第二章 整體系統架構簡介 16 2.1 系統應用情境 16 2.2 系統架構流程 18 第三章 基於生成對抗模型之客製化影像顧慮特徵濾除 20 3.1 決定學習目標 20 3.2 特徵對抗模型訓練 20 3.2.1 模型訓練設計 20 3.2.2 模型訓練損失之評估 22 3.3 建立目標匹配模型 23 3.3.1 設定顧慮議題 23 3.3.2 查詢知識庫 24 3.3.3 建立本體論關係網路 25 3.3.4 計算議題與標籤關聯性 26 第四章 基於關鍵影格之模型即時調適 28 4.1 辨識模型訓練 28 4.2 專家模型之選擇 30 4.3 偵測關鍵影像 31 4.4 後設認知學習策略 32 4.4.1 學什麼 33 4.4.2 何時學 34 4.4.3 如何學 34 第五章 基於智慧合約之可磋商服務品質確保 37 5.1 磋商階段 39 5.2 系統監控階段 40 第六章 實驗結果與討論 42 6.1 基於生成對抗模型之客製化影像顧慮特徵濾除 42 6.1.1 特徵對抗模型實驗 42 6.1.2 特徵對抗模型實驗討論 46 6.1.3 目標比對模型實驗 48 6.1.4 目標比對模型實驗討論 51 6.2 基於關鍵影格之模型即時調適 51 6.2.1 關鍵影格偵測實驗 52 6.2.2 關鍵影格偵測實驗討論 57 6.2.3 模型訓練控制實驗 58 6.2.4 模型訓練控制實驗討論 61 6.3 基於智慧合約之可磋商服務品質確保 62 第七章 結論與未來規劃 65 參考文獻 67 口試委員之建議與回覆 71

    [1] D. Reinsel, J. Gantz, and J. Rydning, "The digitization of the world: from edge to core," ed: November, 2018.
    [2] G. Ananthanarayanan et al., "Real-time video analytics: The killer app for edge computing," Computer, vol. 50, no. 10, pp. 58-67, 2017.
    [3] S. Ojha and S. Sakhare, "Image processing techniques for object tracking in video surveillance-A survey," in 2015 International Conference on Pervasive Computing (ICPC), 2015, pp. 1-6: IEEE.
    [4] S. K. Shah and A. Bedagkar-Gala, "Person Re-identification in Wide Area Camera Networks," in Wide Area Surveillance: Springer, 2014, pp. 137-162.
    [5] X. Wang, "Intelligent multi-camera video surveillance: A review," Pattern Recognition Letters, vol. 34, no. 1, pp. 3-19, 2013/01/01/ 2013.
    [6] K. Kim and L. S. Davis, "Multi-camera tracking and segmentation of occluded people on ground plane using search-guided particle filtering," in European Conference on Computer Vision, 2006, pp. 98-109: Springer.
    [7] J. Zhu, S. Liao, Z. Lei, and S. Z. Li, "Multi-label convolutional neural network based pedestrian attribute classification," Image and Vision Computing, vol. 58, pp. 224-229, 2017.
    [8] Y. Yan, E. Ricci, R. Subramanian, G. Liu, O. Lanz, and N. Sebe, "A multi-task learning framework for head pose estimation under target motion," IEEE transactions on pattern analysis and machine intelligence, vol. 38, no. 6, pp. 1070-1083, 2016.
    [9] H. Li, K. Ota, and M. Dong, "Learning IoT in edge: deep learning for the internet of things with edge computing," IEEE Network, vol. 32, no. 1, pp. 96-101, 2018.
    [10] M. Bilkhu and H. Ayyubi, "Human Activity Recognition for Edge Devices," arXiv preprint arXiv:1903.07563, 2019.
    [11] J. Gama, I. Žliobaitė, A. Bifet, M. Pechenizkiy, and A. Bouchachia, "A survey on concept drift adaptation," ACM computing surveys (CSUR), vol. 46, no. 4, p. 44, 2014.
    [12] P. P. Angelov and X. Gu, "Toward Anthropomorphic Machine Learning," Computer, vol. 51, no. 9, pp. 18-27, 2018.
    [13] C.-H. Lu and C.-H. Yu, "Online Data Stream Analytics for Dynamic Environments Using Self-Regularized Learning Framework," IEEE Systems Journal, 2019.
    [14] O. E. Dictionary, "cognition, n.". Oxford University Press.
    [15] P. Thagard, "Cognitive Science," in The Stanford Encyclopedia of Philosophy, E. N. Zalta, Ed., Spring 2019 ed: Metaphysics Research Lab, Stanford University, 2019.
    [16] B. J. Zimmerman, "Attaining self-regulation: A social cognitive perspective," in Handbook of self-regulation: Elsevier, 2000, pp. 13-39.
    [17] C.-R. Chen and C.-H. Lu, "IoT-enabled Wide Area Monitoring System with Privacy-preserving and Model-adaptable Designs, 2018," in 2018 Taiwan Conference on Software Engineering (TCSE) Conference, Tainan, 2018.
    [18] I. Žliobaitė, A. Bifet, B. Pfahringer, and G. Holmes, "Active learning with drifting streaming data," IEEE transactions on neural networks and learning systems, vol. 25, no. 1, pp. 27-39, 2013.
    [19] M. Pratama, G. Zhang, M. J. Er, and S. Anavatti, "An incremental type-2 meta-cognitive extreme learning machine," IEEE transactions on cybernetics, vol. 47, no. 2, pp. 339-353, 2017.
    [20] M. Pratama, E. Lughofer, J. Lu, M. J. Er, and S. Anavatti, "Data Driven Modelling Based on Recurrent Interval-Valued Metacognitive Scaffolding Fuzzy Neural Network," Neurocomputing, vol. PP, no. 99, pp. 1-44, 2016.
    [21] A. Cavoukian, "Privacy by design: The 7 foundational principles," Information and Privacy Commissioner of Ontario, Canada, vol. 5, 2009.
    [22] A. Alruban, N. Clarke, F. Li, and S. Furnell, "Proactive biometric-enabled forensic imprinting," in 2016 International Conference On Cyber Security And Protection Of Digital Services (Cyber Security), 2016, pp. 1-15: IEEE.
    [23] J. Wang, B. Amos, A. Das, P. Pillai, N. Sadeh, and M. Satyanarayanan, "A scalable and privacy-aware IoT service for live video analytics," in Proceedings of the 8th ACM on Multimedia Systems Conference, 2017, pp. 38-49: ACM.
    [24] A. J. Bose and P. Aarabi, "Adversarial attacks on face detectors using neural net based constrained optimization," in 2018 IEEE 20th International Workshop on Multimedia Signal Processing (MMSP), 2018, pp. 1-6: IEEE.
    [25] N. Foukia, D. Billard, and E. Solana, "PISCES: A framework for privacy by design in IoT," in 2016 14th Annual Conference on Privacy, Security and Trust (PST), 2016, pp. 706-713: IEEE.
    [26] B. Deva, S. R. Garzon, and S. Schünemann, "A context-sensitive privacy-aware
    framework for proactive location-based services," in 2015 9th International Conference on Next Generation Mobile Applications, Services and Technologies, 2015, pp. 138-143: IEEE.
    [27] Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, "Image quality assessment: from error visibility to structural similarity," IEEE transactions on image processing, vol. 13, no. 4, pp. 600-612, 2004.
    [28] X. Zhu, Y. Xiong, J. Dai, L. Yuan, and Y. Wei, "Deep feature flow for video recognition," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017, pp. 2349-2358.
    [29] A. Outchakoucht, E. Hamza, and J. P. Leroy, "Dynamic access control policy based on blockchain and machine learning for the internet of things," Int. J. Adv. Comput. Sci. Appl, vol. 8, no. 7, pp. 417-424, 2017.
    [30] E. C. Ferrer, "The blockchain: a new framework for robotic swarm systems," in Proceedings of the Future Technologies Conference, 2018, pp. 1037-1058: Springer.
    [31] G. J. Mendis, M. Sabounchi, J. Wei, and R. Roche, "Blockchain as a Service: An Autonomous, Privacy Preserving, Decentralized Architecture for Deep Learning," arXiv preprint arXiv:1807.02515, 2018.
    [32] A. Pouraghily, M. N. Islam, S. Kundu, and T. Wolf, "Privacy in Blockchain-Enabled IoT Devices," in 2018 IEEE/ACM Third International Conference on Internet-of-Things Design and Implementation (IoTDI), 2018, pp. 292-293: IEEE.
    [33] H. Li, Y. Li, and F. Porikli, "Deeptrack: Learning discriminative feature representations online for robust visual tracking," IEEE Transactions on Image Processing, vol. 25, no. 4, pp. 1834-1848, 2016.
    [34] S.-T. Chen, C. Cornelius, J. Martin, and D. H. P. Chau, "Shapeshifter: Robust physical adversarial attack on faster r-cnn object detector," in Joint European Conference on Machine Learning and Knowledge Discovery in Databases, 2018, pp. 52-68: Springer.
    [35] N. Papernot, P. McDaniel, I. Goodfellow, S. Jha, Z. B. Celik, and A. Swami, "Practical black-box attacks against machine learning," in Proceedings of the 2017 ACM on Asia conference on computer and communications security, 2017, pp. 506-519: ACM.
    [36] P. Isola, J.-Y. Zhu, T. Zhou, and A. A. Efros, "Image-to-image translation with conditional adversarial networks," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 1125-1134.
    [37] K. Simonyan and A. Zisserman, "Very deep convolutional networks for large-scale image recognition," arXiv preprint arXiv:1409.1556, 2014.
    [38] K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770-778.
    [39] L. Huang, C. Sun, Y. Chi, and H. Xu, "Auto-Construction of Domain Ontology based on Wikipedia and Scientific Papers," Journal of Residuals Science & Technology, vol. 13, no. 6, 2016.
    [40] Y. Wang, T. Peng, and W. Zuo, "Hyponymy Graph Model for Word Semantic Similarity Measurement," Chinese Journal of Electronics, vol. 24, no. 1, pp. 96-101, 2015.
    [41] J. Redmon and A. Farhadi, "Yolov3: An incremental improvement," arXiv preprint arXiv:1804.02767, 2018.
    [42] T.-Y. Lin, P. Dollár, R. Girshick, K. He, B. Hariharan, and S. Belongie, "Feature pyramid networks for object detection," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 2117-2125.
    [43] S. Ren, K. He, R. Girshick, and J. Sun, "Faster r-cnn: Towards real-time object detection with region proposal networks," in Advances in neural information processing systems, 2015, pp. 91-99.
    [44] A. Bogner, M. Chanson, and A. Meeuw, "A decentralised sharing app running a smart contract on the ethereum blockchain," in Proceedings of the 6th International Conference on the Internet of Things, 2016, pp. 177-178: ACM.
    [45] Z. Zhang, Y. Song, and H. Qi, "Age progression/regression by conditional adversarial autoencoder," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017, pp. 5810-5818.
    [46] I. J. Goodfellow, J. Shlens, and C. Szegedy, "Explaining and harnessing adversarial examples," arXiv preprint arXiv:1412.6572, 2014.
    [47] R. Misra, M. Wan, and J. McAuley, "Decomposing fit semantics for product size recommendation in metric spaces," in Proceedings of the 12th ACM Conference on Recommender Systems, 2018, pp. 422-426: ACM.
    [48] Y. Wong, S. Chen, S. Mau, C. Sanderson, and B. C. Lovell, "Patch-based probabilistic image quality assessment for face selection and improved video-based face recognition," in CVPR 2011 WORKSHOPS, 2011, pp. 74-81: IEEE.

    無法下載圖示 全文公開日期 2025/01/16 (校內網路)
    全文公開日期 2025/01/16 (校外網路)
    全文公開日期 2025/01/16 (國家圖書館:臺灣博碩士論文系統)
    QR CODE