以兩階段深度學習網路分析桌球軌跡追蹤與落點偵測

球類運動的自動化追蹤與分析為近年許多研究探討的問題，在競技桌球講求知己知彼，誰能透過人工智慧技術來精準分析桌球比賽中複雜對抗過程，將能有效掌握球小、速度快、旋轉變化多之特性，進而取得致勝先機。本研究目的是透過電腦視覺技術來判斷球體移動的軌跡與落點，使競技桌球之情蒐任務更客觀且更具效率，以利選手從事戰術訓練。以桌球比賽之影片為對象，並提出兩階段式的深度學習網路軌跡預測方法，結合落點偵測演算法來偵測桌球軌跡與落點。第一階段將影像中球體的位置進行偵測，第二階段採用透過第一階段所產生感興趣區域的連續畫面，對桌球軌跡座標進行預測。研究結果顯示：在幀率120 FPS（frames per second，每秒幀數）的影片中，球軌跡追蹤之精準度及召回率分別達到95.51%與92.65%，落點偵測則分別達到90.90%與88.23%；而在使用幀率60 FPS的影片中，分析球軌跡追蹤之精準度及召回率分別達到94.69%與76.80%，在落點偵測則分別達到86.36%與92.59%。本研究結論指出實驗結果證明深度學習方法的有效性，在落點偵測的階段，軌跡座標的準確度及精準度是影響偵測效果的主要因素，因此在優化兩階段式的深度學習網路，讓訓練後的模型提供高準確率與精準度的座標點，以利落點偵測與後續戰略分析之應用。

關鍵字

電腦視覺；演算法；感興趣區域；精準度；召回率

並列摘要

Many studies in recent years have discussed the automatic tracking and analysis of ball sports. In table tennis (TT) competition, it is important to know who uses the artificial intelligence (AI) technology to accurately analyze the complex competition process in TT to manage the characteristics of small-size ball, speed and several rotation changes, leading to a winning opportunity. The purpose of this study was to determine the trajectory and bounce point of the ball movement through computer vision technology. We proposed a two-stage deep learning network algorithm for the ball tracking and bounce estimation from TT videos. The network's first stage detected the TT ball position in the image. In the second stage, we used obtained region of interest (ROI) image sequences from the first stage to predict the precise coordinates of the TT ball. Results showed that the precision and recall of ball tracking reached 95.51% and 92.65%, and the bounce detection reached 90.90% and 88.23%, respectively, when the film with a frame rate of 120 FPS (frames per second) was used. In the film with a frame rate of 60 FPS, the ball tracking precision and recall reached 94.69% and 76.80%, and the bounce detection reached 86.36% and 92.59%, respectively. These findings demonstrated the effectiveness of the proposed deep learning methods in this study. In bounce calculation, the accuracy and precision of trajectory were the main factors that influence the calculation effect. Therefore, we optimized the two-stage degree learning networks to provide the trained model with high accuracy and precision coordinates to facilitate the application of bounce detection and subsequent strategic analysis.

並列關鍵字

computer vision ； algorithm ； region of interest ； precision ； recall

參考文獻

王威堯、張凱翔、陳霆峰、王志全、彭文志、易志偉（2020）。Badminton Coach AI：基於深度學習之羽球賽事資訊分析平台。體育學報，53(2)，201-213。https://doi.org/10.6222/pej.202 006_53(2).0005 [Wang, W.-Y., Chang, K.-S., Chen, T.-F., Wang, C.-C., Peng, W.-C., & Yi, C.-W. (2020). Badminton Coach AI: A badminton match data analysis platform based on deep learning. Physical Education Journal, 53(2), 201-213. https://doi.org/10.6222/pej.202006_53(2).0005]

相子元、石又、何金山（2012）。感測科技於運動健康科學之應用。體育學報，45(1)，1-12。https://doi.org/10.6222/pej.4501.201203.0701 [Shiang, T.-Y., Shih, Y., & Ho, C.-S. (2012). the applications of sensor technology for exercise and sport science. Physical Education Journal, 45(1), 1-12. https://doi.org/10.6222/pej.4501.201203.0701]

許銘華、于承煥、許英麟、蔡亞倫、吳昇光（2022）。以 3S 理論與機器學習分析頂尖混雙桌球選手接發球技戰術：林昀儒／鄭怡靜之個案研究。大專體育學刊，24(4)，563-583。https://doi.org/1 0.5297/ser.202212_24(4).0009 [Hsu, M.-H., Yu, C.-H., Hsu, Y.-L., Tsai, Y.-L., & Wu, S.-K. (2022). Analysis of receiving techniques and tactics of top mixed doubles table tennis players based on 3S principle and machine learning: a case study of Yun-Ju Lin/I-Ching Cheng. Sports & Exercise Research, 24(4), 563-583. https://doi.org/10.5297/ser.202212_24(4).0009]

熊志超、周資眾、許銘華（2020）。40+ 競技桌球銜接技術對高水準運動員戰術發揮的影響。中華體育季刊，34(4)，273-285。https://doi.org/1 0.6223/qcpe.202012_34(4).0006 [Hung, C.-C., Chou, T.-C., & Hsu, M.-H. (2020). the impact of 40+ competitive table tennis connecting techniques on the tactics of high-level athletes. Quarterly of Chinese Physical Education, 34(4), 273-285. https://doi.org/10.6223/qcpe.202012_34(4).0006]

Ashford, D., Bennett, S. J., & Davids, K. (2006). Observational modeling effects for movement dynamics and movement outcome measures across differing task constraints: A meta-analysis. Journal of Motor Behavior, 38(3), 185-205. https://doi.org/10.3200/JMBR.38.3.185-205

主題瀏覽