臺灣博碩士論文加值系統

基於電腦視覺的人工智能運動分析技術，不僅迅速有效的自動追蹤球員軌跡、收集賽事數據，更能解決傳統人為判斷與識別上過於主觀的問題。近年來，已經發展為主流運動賽事分析與應用的創新商業模式。然而，傳統的數據紀錄方式仰賴人工作業，這對於時間與成本上造成相當的負擔。
因此，本研究致力於研發數據產生系統，藉由電腦視覺在影像處理與分析上的優勢，於賽事影像上進行識別作業，紀錄球員在球場上的表現，通過仿射轉換的方法將影像資訊投影至平面圖，用來產生賽事中球員的絕對座標、奔跑速度、移動距離等資訊，該系統架構能夠有效地靈活化整體作業之流程，達成自動化數據紀錄之成效。

Based on computer Vision, the artificial intelligent motion analysis technology not only quickly and effectively track the player’s trajectory automatically, but also collect event data. It can solve the problem of tradition subjective judgment and recognition. In recent years, it has developed into an innovative business model for the analysis and application of mainstream sports events. However, the traditional method of data recording relies on manual work, which imposes a considerable burden on time and cost.
Therefore, this research is devoted to the development of dataset generation system. Through the advantages of computer vision and image processing in detection and recognition, Identifying the game on the broadcast sport video and recording the performance of the player of the court. We use the affine transformation method to project game image information to the floor plan to generate the absolute coordinates, running speed, moving distance and other information of the players in the events. This system architecture can effectively flex the overall operation process and achieve the results of automated data records.

摘 要 I
Abstract II
誌 謝 III
第一章 緒論 1
1.1 研究動機與背景 1
1.2 研究動機與目的 4
1.3 章節概要 5
第二章 文獻探討 6
2.1物件偵測(Object Detection) 6
2.1.1 區域檢測進程介紹 6
2.1.2滑動窗口檢測(Sliding window detector) 7
2.1.3錨點偵測(Anchor detector) 7
2.1.4方向梯度直方圖(Anchor detector) 8
2.1.4深度神經網路(Deep Neural Network) 9
2.2區域快速卷積神經網路(Faster Region-based Convolution Neural Network) 10
2.3殘差網路(Residual Network) 12
2.4霍夫線偵測(Hough Line Detect) 13
2.5仿射轉換(Affine Transform) 14
第三章 研究方法 15
3.1 方法流程圖 15
3.1.1 流程步驟 15
3.1.2 資料選擇 15
3.1.3 資料處理 16
3.1.4 資料轉換 17
3.1.5 電腦視覺及座標映射 18
3.2 模型訓練調配 18
3.2.1 Faster R-CNN 錨點參數調配 18
3.2.2 ResNet識別網路架構 19
3.3自動化數據紀錄 20
3.3.1 球場追蹤 20
3.3.2座標映射 24
第四章 結果討論分析 28
4.1 NBA球場廣播視頻 28
4.2 Faster R-CNN分類結果 28
4.2 ResNet識別結果 29
4.3 仿射轉換結果討論 30
4.4 實時運行時間 32
第五章 結論與未來展望 33
參考文獻 34

[1] S. Dev, B. Wen, Y. H. Lee, and S. Winkler, “Ground-based image analysis: A tutorial on machine-learning techniques and applications,” IEEE Geoscience and Remote Sensing Magazine, vol. 4, no. 2, pp. 79-93, 2016.
[2] J. Ding, X. H. Hu, and V. Gudivada, “A Machine Learning Based Framework for Verification and Validation of Massive Scale Image Data,” IEEE Transactions on Big Data, vol. PP, no. 9, pp. 1-1, 2017.
[3] S. B. Jemaa, M. Hammami, and H. Ben-Abdallah, “Data-mining process: application for hand detection in contact free settings,” IET Image Processing, vol. 7, no. 8, pp. 742-750, 2013.
[4] S. Jia, J. Hu, J. Zhu, X. Jia, and Q. Li, “Three-Dimensional Local Binary Patterns for Hyperspectral Imagery Classification,” IEEE Transactions on Geoscience and Remote Sensing, vol. 55, no. 4, pp. 2399–2413, April 2017.
[5] L. Essannouni, and D. Aboutajdine, “Correlation of robust Haar-like feature,” Electronics Letters, vol. 47, no. 17, pp. 961–962, August 2011.
[6] Y. Wang, L. Du, and H. Dai, “Unsupervised SAR Image Change Detection Based on SIFT Keypoints and Region Information,” IEEE Geoscience and Remote Sensing Letters, vol. 13, no. 7, pp. 931–935, July 2016.
[7] S. Tuermer, F. Kurz, P. Reinartz, U. Stilla, “Airborne Vehicle Detection in Dense Urban Areas Using HOG Features and Disparity Maps,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 6, no. 6, pp. 2327–2337, December 2013.
[8] Y. Lecun, L. Bottou, Y. Bengio, and P. Haffner, “Gradient-based learning applied to document recognition,” Proceedings of the IEEE, vol. 86, no. 11, pp. 2278-2324, 1998. 
[9] A. Krizhevsky, I. Sutskever, and G. E. Hinton, “ImageNet Classification with Deep Convolutional Neural Networks,” Advances in Neural Information Processing Systems, vol. 25, no. 2, pp. 1097-1105, 2012.
[10] Y.-S. Zhao, J. Wu, Y.-F. Zhu, H.-X. Yu, and R. Xiong, “A Learning Framework Towards Real-time Detection and Localization of a Ball for Robotic Table Tennis System,” IEEE International Conference on Real-time Computing and Robotics, pp. 97–102, July 2017.
[11] P. Felzenszwalb, R. Girshick, D. McAllester, and D. Ramanan. Object detection with discriminatively trained part based models. TPAMI, 2010
[12] W.-L. Lu, J.-A. Ting, J. J. Little, and K. P. Murphy, “Learning to track and identify players from broadcast sports videos,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 35, no. 7, pp. 1704–1716, 2013.
[13] W.-L. Lu, J.-A. Ting, K. P. Murphy, and J. J. Little, “Identifying Players in Broadcast Sports Videos using Conditional Random Fields,” in CVPR, 2011.
[14] P.-C. Wen, W.-C. Cheng, Y.-S. Wang, H.-K. Chu, N. C. Tang, H.-Y. M. Liao, “Court Reconstruction for Camera Calibration in Broadcast Basketball Video”, IEEE Transactions on Visualization and Computer Graphics, Vol: 22, May 2016
[15] K. Okuma, A. Taleghani, N. de Freitas, J. J. Little, and D. G. Lowe, “A Boosted Particle Filter: Multitarget Detection and Tracking,” in ECCV, 2004.
[16] J. Han, D. Farin, and P. H. N. de With, “A real-time augmented-reality system for sports broadcast video enhancement,” in International Conference on Multimedia, 2007, pp. 337–340.
[17] K. Kumar, P. Parisot, and C. De Vleeschouwer, “Demo: Spatio- temporal template matching for ball detection,” in ACM/IEEE International Conference on Distributed Smart Cameras, 2011, pp. 1–2. 
[18] C. Verleysen and C. D. Vleeschouwer, “Recognition of sport players’ numbers using fast-color segmentation,” in IS&T/SPIE | Electronic Imaging, 2012.
[19] N. Dalal and B. Triggs. “Histograms of Oriented Gradients for Human Detection,” in CVPR, 2005.
[20] C. L. Zitnick and P. Dollar. “Edge Boxes: Locating Object Proposals from Edges.” In ECCV 2014.
[21] C. P. Papageorgiou, M. Oren, and T. Poggio. “A general framework for object detection.” In ICCV 1998.
[22] D. Cervone , A. D’Amour , L. Bornn , and K. Goldsberry, “POINTWISE: Predicting Points and Valuing Decisions in Real Time with NBA Optical Tracking Data”, in MIT SLOAN 2014.
[23] Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun, “Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks” in Neural Information Processing Systems (NIPS) 2015.
[24] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun “Deep Residual Learning for Image Recognition” in CVPR 2015.