臺灣博碩士論文加值系統

本論文設計一叢集式影片物件辨識平台，藉由NVidia 公司所推出的TX1嵌入式開發板所提供的圖形處理器(GPU)進行OpenCV影像辨識，並使用Python的分散式套件dispy在多台機器上進行分散運算。因圖形處理器其結構相異於傳統中央處理器(CPU)，對於多媒體資料具有相當高的運算效率。藉由dispy套件的串聯，使得多台機器能夠同時進行平行運算。
本論文同時設計針對辨識結果進行權重統計之方法，以符合人類在主觀意識上對於影片內容之想法，並進行不同影格之間之相似比對，用於過濾相似影格，節省運算總量，最後對於進行加速時所造成的效能損失進行探討。
本論文實作之平台，在無失真之情況下相較於傳統辨識方法可以獲得12倍的效能提升，在允許部分失真之情況下，忽略 20%影格差異能達到約45倍效能提升，忽略 40%影格差異能達到約85倍效能提升，大幅度降低多重物件辨識所需時間，使得相關領域能夠大幅度節省辨識所花費的時間。

Designing an object detection platform cluster by using Jetson TX1, which developed by NVidia. Jetson TX1 provides an environment with Graphics processing unit(GPU) supporting CUDA. By the GPU we can process OpenCV object detection faster than CPU. We can also connect these development kits by using Python extension – dispy, and make development kits parallel compute the data. GPU is quite different than CPU, its’ multi-core can process media data with high-performance.
We also design an algorithm to calculate detection result, according to its’ frequency and multiplied by weights. Filtering similar frame can also reduce process time. And finally make a discussion about performance loss.
The platform we make can get 12 time faster than traditional detection. Ignoring 20% of frame difference can get 45 time faster. Ignoring 40% of frame difference can get 85 time faster. By our experiment, this platform can significantly reduce detection time.

致謝辭 I
摘要 II
Abstract III
目錄 IV
圖目錄 VII
表目錄 IX
第1章 緒論 1
1.1 動機 1
1.2 研究目的 1
1.3 論文架構 1
第2章 相關研究探討 3
2.1 影片內容分析 3
2.1.1 背景相減法與區塊分析應用於圖像辨識 3
2.1.2 SVM物件辨識 3
2.1.3 哈爾分類器物件辨識 4
2.2 叢集式運算架構 5
2.2.1 貝奧武夫機群 5
2.2.2 訊息傳遞介面 6
2.3 圖形處理晶片 6
第3章 叢集式圖像處理器影片物件辨識系統設計方法 7
3.1 分配與運算演算法 7
3.2 影片內容關聯程度 12
3.3 畫面相似程度比對 13
3.4 運算效能損失評估 14
第4章 系統實現與實驗結果 16
4.1 實驗環境 16
4.1.1 實驗硬體環境 16
4.1.2 實驗軟體環境 17
4.1.3 OpenCV 18
4.1.4 CUDA 18
4.1.5 dispy叢集式運算套件 19
4.1.6 影片物件辨識相關檔案 23
4.1.7 OpenCV Python Extension 26
4.1.8 運算節點狀態管理 26
4.2 實驗方法 27
4.3 單機運算比較 28
4.4 叢集式運算比較 30
4.4.1 節點數量與網路環境實驗 30
4.4.2 單一工作傳送複數影格實驗 32
4.4.3 相似畫面過濾 36
4.4.4 各實驗辨識速率整理 37
第5章 結論與未來展望 38
5.1 結論 38
5.2 未來展望 38
第6章 參考文獻 39

[1] S. Dasiopoulou, V. Mezaris, I. Kompatsiaris, V. K. Papastathis and M. G. Strintzis, "Knowledge-assisted semantic video object detection," in IEEE Transactions on Circuits and Systems for Video Technology, vol. 15, no. 10, pp. 1210-1224, Oct. 2005.
[2] S. Nadimi and B. Bhanu, "Physical models for moving shadow and object detection in video," in IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 26, no. 8, pp. 1079-1087, Aug. 2004.
[3] Vapnik, Vladimir. The nature of statistical learning theory. Springer Science & Business Media, 2013.
[4] Vapnik, V.: Invited Speaker. IPMU Information Processing and Management 2014)
[5] Barghout, Lauren. "Spatial-Taxon Information Granules as Used in Iterative Fuzzy-Decision-Making for Image Segmentation." Granular Computing and Decision-Making. Springer International Publishing, 2015. 285-318.
[6] P. Viola and M. Jones, "Rapid object detection using a boosted cascade of simple features," Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001, 2001, pp. I-511-I-518 vol.1.
[7] K. Mikolajczyk, B. Leibe and B. Schiele, "Multiple Object Class Detection with a Generative Model," 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06), 2006, pp. 26-36.
[8] D. Shin, I. Hong and H. J. Yoo, "An 1.92mW Feature Reuse Engine based on inter-frame similarity for low-power object recognition in video frames," 2014 IEEE International Symposium on Circuits and Systems (ISCAS), Melbourne VIC, 2014, pp. 758-761.
[9] Huiyuan Fu, H. Ma and A. Ming, "EGMM: An enhanced Gaussian mixture model for detecting moving objects with intermittent stops," 2011 IEEE International Conference on Multimedia and Expo, Barcelona, 2011, pp. 1-6.
[10] C. Stauffer and W. E. L. Grimson, "Adaptive background mixture models for real-time tracking," Proceedings. 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No PR00149), Fort Collins, CO, 1999, pp. 252 Vol. 2.
[11] P. K. Bhaskar and S. P. Yong, "Image processing based vehicle detection and tracking method," 2014 International Conference on Computer and Information Sciences (ICCOINS), Kuala Lumpur, 2014, pp. 1-5.
[12] Hong Lu, Yingbin Zheng, Xiangyang Xue and Yuejie Zhang, "Content and context-based multi-label image annotation," 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, Miami, FL, 2009, pp. 61-68.
[13] W. Zhu, C. Luo, J. Wang and S. Li, "Multimedia Cloud Computing," in IEEE Signal Processing Magazine, vol. 28, no. 3, pp. 59-69, May 2011.
doi: 10.1109/MSP.2011.940269
[14] Zhe Fan, Feng Qiu, A. Kaufman and S. Yoakum-Stover, "GPU Cluster for High Performance Computing," Supercomputing, 2004. Proceedings of the ACM/IEEE SC2004 Conference, 2004, pp. 47-47.
[15] OpenCV CUDA supported, http://opencv.org/platforms/cuda.html
[16] NVidia embedded-systems-dev-kits-modules, http://www.nvidia.com.tw/object/embedded-systems-dev-kits-modules-tw.html
[17] CUDA parallel computing, http://docs.nvidia.com/cuda/parallel-thread-execution/index.html
[18] ImageNet image database, http://image-net.org/
[19] L4T 24.2.1 official document, https://developer.nvidia.com/embedded/dlc/l4t-documentation-24-2-1
[20] Zhou Yang, Xiqiao Lin, Wenqian Jiang and Gang Li, "An electricity data cluster analysis method based on SAGA-FCM algorithm," 2017 IEEE 14th International Conference on Networking, Sensing and Control (ICNSC), Calabria, Italy, 2017, pp. 389-393.
[21] N. M. F. Qureshi, D. R. Shin, I. F. Siddiqui and B. S. Chowdhry, "Storage-Tag-Aware Scheduler for Hadoop Cluster," in IEEE Access, vol. 5, no. , pp. 13742-13755, 2017.
[22] M. Jasiński, T. Sikorski, J. Karpiński and M. Zenger, "Cluster analisis of long-term power quality data," 2016 Electric Power Networks (EPNet), Szklarska Poreba, Poland, 2016, pp. 1-6.
[23] V. A. Raikhlin and R. K. Klassen, "Can GPU-accelerator significantly increase the effectiveness of conservative DBMS considerable volumes on cluster platforms?," 2017 International Siberian Conference on Control and Communications (SIBCON), Astana, Kazakhstan, 2017, pp. 1-5.
[24] Raikhlin V.A. “Simulation of Distributed Database Machines”. Programming and Computer Software. Vol. 22, Issue 2, 1996, P. 68-74
[25] W. Ai, K. Li, C. Chen, J. Peng and K. Li, "DHCRF: A Distributed Conditional Random Field Algorithm on a Heterogeneous CPU-GPU Cluster for Big Data," 2017 IEEE 37th International Conference on Distributed Computing Systems (ICDCS), Atlanta, GA, USA, 2017, pp. 2372-2379.
[26] M. Krajecki, J. Loiseau, F. Alin and C. Jaillet, "BFS Traversal on Multi-GPU Cluster," 2016 IEEE Intl Conference on Computational Science and Engineering (CSE) and IEEE Intl Conference on Embedded and Ubiquitous Computing (EUC) and 15th Intl Symposium on Distributed Computing and Applications for Business Engineering (DCABES), Paris, France, 2016, pp. 594-599.
[27] Kuan-Yu Yeh, Hui-Jun Cheng, Jin Ye, Jyh-Da Wei and Chun-Yuan Lin, "Constructing a GPU cluster platform based on multiple NVIDIA Jetson TK1," 2016 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Shenzhen, 2016, pp. 917-922.
[28] P. K. Mohanty, M. Reza, P. Kumar and P. Kumar, "Implementation of Cubic Spline Interpolation on Parallel Skeleton Using Pipeline Model on CPU-GPU Cluster," 2016 IEEE 6th International Conference on Advanced Computing (IACC), Bhimavaram, 2016, pp. 747-751.
[29] S. Aslan, H. Badem, D. Karaboga, A. Basturk and T. Ozcan, "Accelerating Discrete Haar Wavelet Transform on GPU cluster," 2015 9th International Conference on Electrical and Electronics Engineering (ELECO), Bursa, 2015, pp. 1237-1240.
[30] M. Sourouri, J. Langguth, F. Spiga, S. B. Baden and X. Cai, "CPU+GPU Programming of Stencil Computations for Resource-Efficient Use of GPU Clusters," 2015 IEEE 18th International Conference on Computational Science and Engineering, Porto, 2015, pp. 17-26.
[31] I. Savran and J. R. Rose, "HPMA: High-performance metagenomic alignment tool, on a large-scale GPU cluster," 2015 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Washington, DC, 2015, pp. 629-634.
[32] K. Hamidouche, A. Venkatesh, A. A. Awan, H. Subramoni, C. H. Chu and D. K. Panda, "Exploiting GPUDirect RDMA in Designing High Performance OpenSHMEM for NVIDIA GPU Clusters," 2015 IEEE International Conference on Cluster Computing, Chicago, IL, 2015, pp. 78-87.
[33] J. Zhang, S. You and L. Gruenwald, "Tiny GPU Cluster for Big Spatial Data: A Preliminary Performance Evaluation," 2015 IEEE 35th International Conference on Distributed Computing Systems Workshops, Columbus, OH, 2015, pp. 142-147.
[34] R. Shi et al., "Designing efficient small message transfer mechanism for inter-node MPI communication on InfiniBand GPU clusters," 2014 21st International Conference on High Performance Computing (HiPC), Dona Paula, 2014, pp. 1-10.
[35] S. K. Das, S. G. and V. Ashok, "Learning Based Performance and Power Efficient Cluster Resource Manager for CPU-GPU Cluster," 2014 Fourth International Conference of Emerging Applications of Information Technology, Kolkata, 2014, pp. 161-166.