全世界打棒壘球的人數非常的多，而在棒壘球中，打擊是一項基本的動作，但它同時也是最難掌握的技巧之一。打擊動作需要不斷的練習和適當的指導。經驗豐富的教練能夠為棒壘球員提供即時的動作診斷和反饋，從而縮短學習曲線。可惜的是，不是每位球員都有一位經驗豐富的教練能夠從旁指導。過去在虛擬運動教練的開發中，有些需要昂貴且笨重的動作追蹤設備，有些則是沒辦法提供有效的動作診斷及指導。本文針對棒壘球打擊練習的揮棒動作提出了一個低成本動作診斷及指導的虛擬教練。這個虛擬教練只需要一個穿戴式裝置及一台相機，例如智慧型手機的相機，來捕捉棒壘球員的動作。收集到的數據會加以分析並得到球員在不同揮棒階段的動作特徵，然後將其與專家們演示的參考動作進行比較以便檢測可能的動作錯誤並向球員提供適合的指導。我們收集了4位專家和8位初學者球員，總共約1200次的打擊揮棒動作。虛擬教練的評估結果也會與指導經驗豐富的教練做比較。實驗證實，我們的虛擬教練與人為指導相比，在揮棒動作的診斷及指導方面能夠達到約80%的準確度。

The population of playing baseball/softball worldwide is quite large. In baseball/softball, batting is a fundamental action, but it is also one of the most difficult skills to master. Batting requires constant practice and proper guidance. Experienced coaches can provide instant diagnoses and feedback tailored for individual players, whereby shortening the learning curve. Unfortunately, not every player can have an experienced coach by the side. Past efforts on coaching tools for sports are either too expensive and awkward to use, or too limited to provide useful diagnosis and guidance. In this thesis, we introduce a low-cost diagnosis and guidance tool for batting practice in baseball/softball. The tool requires only one wearable device and one camera, such as the one on the smartphone, to capture the player's motion. The collected data are analyzed to derive the features of the player's actions in different swing stages, which are then compared with the reference actions to detect possible mistakes and provide suitable guidance to the player. We have collected about 1200 swing motions from four experts and eight novice players. The evaluation results from our tool are compared with those from the experienced coach. The experiments show that our tool can achieve about 80% accuracy in terms of diagnosis and guidance.

[1] http://www.zepp.com/en-us/baseball/.
[2] https://buy.garmin.com/en-US/US/p/579018.
[3] https://smashprosports.com/.
[4] D. Hirayama, K. Yoshizawa, H. Sogo, and T. Henmi. Quantitative comparison of technical differences in baseball batting motion by motion analysis. In 2016 International Conference on Advanced Mechatronic Systems (ICAMechS), pages 115–120, Nov 2016.
[5] Brittany Dowling and Glenn S. Fleisig. Kinematic comparison of baseball batting off of a tee among various competition levels. Sports Biomechanics, 15(3):255–269, 2016. PMID: 27278749.
[6] Dharmayanti, M. Iqbal, A. Suhendra, and A. Benny Mutiara. Velocity and acceleration analysis from kinematics linear punch using optical motion capture. In 2017 Second International Conference on Informatics and Computing (ICIC), pages 1–6, Nov 2017.
[7] K. Kolykhalova, A. Camurri, G. Vlpe, M. Sanguineti, E. Puppo, and R. Niewiadomski. A multimodal dataset for the analysis of movement qualities in karate martial art. In 2015 7th International Conference on Intelligent Technologies for Interactive Entertainment (INTETAIN), pages 74–78, June 2015.
[8] Zhe Cao, Tomas Simon, Shih-En Wei, and Yaser Sheikh. Realtime multi-person 2d pose estimation using part affinity fields. In CVPR, 2017.
[9] D. Leightley, J. S. McPhee, and M. H. Yap. Automated analysis and quantification of human mobility using a depth sensor. IEEE Journal of Biomedical and Health Informatics, 21(4):939–948, July 2017.
[10] Fotini Patrona, Anargyros Chatzitofis, Dimitrios Zarpalas, and Petros Daras. Motion analysis: Action detection, recognition and evaluation based on motion capture data. Pattern Recognition, 76:612 – 622, 2018.
[11] P. Parmar and B. T. Morris. Learning to score olympic events. In 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), volume 00, pages 76–84, July 2017.
[12] S. Qiao, Y. Wang, and J. Li. Real-time human gesture grading based on open- pose. In 2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI), pages 1–6, Oct 2017.
[13] A. Ahmadi, E. Mitchell, F. Destelle, M. Gowing, N. E. OConnor, C. Richter, and K. Moran. Automatic activity classification and movement assessment during a sports training session using wearable inertial sensors. In 2014 11th International Conference on Wearable and Implantable Body Sensor Networks, pages 98–103, June 2014.
[14] M. Sharma, R. Srivastava, A. Anand, D. Prakash, and L. Kaligounder. Wearable motion sensor based phasic analysis of tennis serve for performance feedback. In 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 5945–5949, March 2017.
[15] H. Ghasemzadeh and R. Jafari. Coordination analysis of human movements with body sensor networks: A signal processing model to evaluate baseball swings. IEEE Sensors Journal, 11(3):603–610, March 2011.
[16] Doo Young Kwon and Markus Gross. Combining body sensors and visual sensors for motion training. In Proceedings of the 2005 ACM SIGCHI International Conference on Advances in Computer Entertainment Technology, ACE ’05, pages 94–101, New York, NY, USA, 2005. ACM.
[17] Hiroki Nakata, Akito Miura, Michiko Yoshie, and Kazutoshi Kudo. Electromyographic activity of lower limbs to stop baseball batting. Journal of strength and conditioning research, 26 6:1461–8, 2012.
[18] Rob Gray. A model of motor inhibition for a complex skill: baseball batting. Journal of experimental psychology. Applied, 15 2:91–105, 2009.
[19] Edward S. Chang, Meghan E. Bishop, Dylan K. Baker, and Robin Vereeke West. Interval throwing and hitting programs in baseball: Biomechanics and rehabilitation. American journal of orthopedics, 45 3:157–62, 2016.
[20] Tomas Simon, Hanbyul Joo, Iain Matthews, and Yaser Sheikh. Hand keypoint detection in single images using multiview bootstrapping. In CVPR, 2017.
[21] Shih-En Wei, Varun Ramakrishna, Takeo Kanade, and Yaser Sheikh. Convolutional pose machines. In CVPR, 2016.
[22] Disheng Yang, Jian Tang, Yang Huang, Chao Xu, Jinyang Li, Liang Hu, Guobin Shen, Chieh-Jan Mike Liang, and Hengchang Liu. Tennismaster: An imu- based online serve performance evaluation system. In Proceedings of the 8th Augmented Human International Conference, AH ’17, pages 17:1–17:8, New York, NY, USA, 2017. ACM.
[23] L. R. Rabiner. A tutorial on hidden markov models and selected applications in speech recognition. Proceedings of the IEEE, 77(2):257–286, Feb 1989.
[24] Lawrence R. Rabiner and B. H. Juang. An introduction to hidden markov models. IEEE ASSP Magazine, 3:4–16, 1986.
[25] K. Liu, C. Chen, R. Jafari, and N. Kehtarnavaz. Fusion of inertial and depth sensor data for robust hand gesture recognition. IEEE Sensors Journal, 14(6):1898–1903, June 2014.
[26] Alessia Saggese, Nicola Strisciuglio, Mario Vento, and Nicolai Petkov. Action recognition by learning pose representations. CoRR, abs/1708.00672, 2017.