以階層式Boosting演算法為基礎之棒球軌跡辨識

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：139

、訪客IP：3.145.42.94

姓名

傅元威(Yuan-Wei Fu) 查詢紙本館藏

畢業系所

通訊工程學系

論文名稱

以階層式Boosting演算法為基礎之棒球軌跡辨識
(Baseball Pitch Recognition UsingHierarchical Boosting Algorithm)

相關論文

★ 應用於車內視訊之光線適應性視訊壓縮編碼器設計	★ 以粒子濾波法為基礎之改良式頭部追蹤系統
★ 應用於空間與CGS可調性視訊編碼器之快速模式決策演算法	★ 應用於人臉表情辨識之強健式主動外觀模型搜尋演算法
★ 結合Epipolar Geometry為基礎之視角間預測與快速畫面間預測方向決策之多視角視訊編碼	★ 基於改良式可信度傳遞於同質區域之立體視覺匹配演算法
★ 多視角視訊編碼之快速參考畫面方向決策	★ 以線上統計為基礎應用於CGS可調式編碼器之快速模式決策
★ 適用於唇形辨識之改良式主動形狀模型匹配演算法	★ 以運動補償模型為基礎之移動式平台物件追蹤
★ 基於匹配代價之非對稱式立體匹配遮蔽偵測	★ 以動量為基礎之快速多視角視訊編碼模式決策
★ 應用於地點影像辨識之快速局部L-SVMs群體分類器	★ 以高品質合成視角為導向之快速深度視訊編碼模式決策
★ 以運動補償模型為基礎之移動式相機多物件追蹤	★ 基於匹配代價曲線特徵之遮蔽偵測之研究

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

[檢視]

[下載]

本電子論文使用權限為同意立即開放。
已達開放權限電子全文僅授權使用者為學術研究之目的，進行個人非營利性質之檢索、閱讀、列印。
請遵守中華民國著作權法之相關規定，切勿任意重製、散佈、改作、轉貼、播送，以免觸法。

摘要(中)

隨著科技的進步，觀眾對於運動轉播影片所能提供額外的娛樂功能也益增，以棒球轉播來說，從單純的在轉播比賽中主播口述並邀請球評講解比賽中的細節，演進至影像中提供球速資訊，甚至更進階需要擁有專業知識的資訊如棒球中的配球理論，這些一般大眾較少能夠接觸到的球賽資訊，也是觀眾較為感興趣的部分。因此本論文以棒球投手球種辨識為目標，整合棒球軌跡追蹤以及球路辨識。此外，並提出以階層式boosting為基礎之球路辨識演算法，取出棒球飛行軌跡特徵後，以其Adaboost分類器延伸之應用於多類別的SAMME演算法選出有效特徵，而弱分類器的設計我們採用以多類別貝式分類法，最後再以這些被賦予不同權重的特徵代表的弱分類器組合成一球路辨識之強分類器。實驗結果顯示，對於多種不同的球場影片，我們的球種辨識正確率平均可以接近80%。

摘要(英)

With advances in technology, the audiences expect additional entertainments of sport broadcasts. To baseball broadcasting, the pure broadcast game only with anchor oral and ball comment discuss the details of the game, go on appear the speed information on screen, even provide more advanced information that audiences have a few opportunities to understand, like pitching tactics theory, this part was attracted to the audiences, too. In this paper, we integrate the pitch recognition with trajectory tracking. Moreover, we propose the hierarchical boosting based pitch. The strong learner adopts the SAMME algorithm that is extended from Adaboost, the weak learner is designed based on the multiclass Bayesian classifier. Our experimental results show that accuracy of pitch recognition could be reached near 80%.

關鍵字(中)

★ 球路辨識
★ adaboost
★ 階層式
★ 棒球

關鍵字(英)

★ recognition
★ pitch
★ baseball
★ adaboost

論文目次

摘要...............................................................................................................................I
Abstract........................................................................................................................II
目錄.............................................................................................................................III
圖目錄..........................................................................................................................V
表目錄......................................................................................................................VIII
第一章緒論..................................................................................................................1
1.1 前言.....................................................................................................................1
1.2 研究動機.............................................................................................................1
1.3 研究方法.............................................................................................................2
1.4 論文架構.............................................................................................................3
第二章棒球軌跡追蹤演算法簡介..............................................................................4
2.1 應用於電視轉播的軌跡重建系統.....................................................................4
2.1.1鷹眼﹝HAWK-EYE﹞.................................................................................4
2.1.2 K-Zone..........................................................................................................5
2.2 球候選點過濾.....................................................................................................6
2.3球軌跡過濾..........................................................................................................7
2.3.1以垂直和水平方向運動為基礎的球候選軌跡過濾...................................7
2.3.2以Kalman filter為基礎的球候選軌跡過濾..............................................10
2.3.3濾出合理的球軌跡.....................................................................................13
2.4總結....................................................................................................................14
第三章球路辨識演算法............................................................................................15
3.1以球路軌跡為主的辨識演算法........................................................................15
3.1.1特徵選擇.............................................................................................15
3.1.2快速球、曲球與其他球種之辨識演算法.........................................16
3.1.3所有常見球種的辨識演算法.............................................................18
3.2 總結...................................................................................................................21
第四章本論文所提出之棒球投手球路追蹤與球種辨識系統................................22
4.1棒球投手球路追蹤演算法................................................................................22
4.2本論文提出之階層式球路辨識演算法............................................................25
4.2.1 Adaboost分類器.........................................................................................25
4.2.2 SAMME介紹..............................................................................................26
4.2.3 Adaboost的弱分類器之設計.....................................................................27
4.2.4貝氏分類法簡介.........................................................................................28
4.2.5階層式球路辨識演算法.............................................................................29
4.3 總結...................................................................................................................32
第五章實驗結果........................................................................................................33
5.1電腦配備與測試影片........................................................................................33
5.2投手球路軌跡追蹤結果....................................................................................34
5.3球種辨識結果....................................................................................................44
5.4比較與討論........................................................................................................48
第六章結論與未來與展望........................................................................................50
參考文獻....................................................................................................................51

參考文獻

[1] H.-T. Chen, H.-S. Chen, M.-H. Hsiao, W.-J. Tsai, and S.-Y. Lee, “A trajectory based ball tracking framework with visual enrichment for broadcast baseball videos,” Journal of Information Science and Engineering 24, pp. 143-157, 2008.
[2] W.-T. Chu, C.-W. Wang, and J.-L. Wu, “Extraction of baseball trajectory and physics based validation for single-view baseball video sequences,” IEEE International Conference on Multimedia & Expo, pp. 1813-1816, 2006.
[3] J.-S. Lee and J.-W. Lin, Baseball pitch recognition for broadcast television, Department of Computer Science Information Engineering, National Chung Cheng University, 2006.
[4] André Guéziec, “Tracking pitches for broadcast television, ” IEEE International Conference on Computer, Vol. 35, No. 3, pp. 38-43, 2002.
[5] N. Owens, C. Harris, and C. Stennett, ”Hawk-eye tennis system,” IEEE International Conference on Visual Information Engineering, pp.182-185, 2003.
[6] G. Welch and G. Bishop, An introduction to the Kalman filter, Department of Computer Science University of North Carolina at Chapel Hill, 2006.
[7] G. Provan, P. Langley, and P. Smyth, “Bayesian network classifiers ,” Journal of Machine Learning, Vol.29, pp. 131-163, 1997.
[8] H. Shun and T. Komura, ”A spatiotemporal approach to extract the 3D trajectory of the baseball from a single view video sequence,” IEEE International Conference on Multimedia and Expo (ICME), pp.1583-1586, 2004.
[9] P. Chang, M. Han, and Y. Gong, “Extract highlights from baseball game video with hidden markov models,“ IEEE International Conference on Image Processing, pp. 609-612, 2002.
[10] A. Björck, Numerical methods for least squares problems, SIAM books.
[11] Y. Freund and R. E. Schapire, “A short introduction to boosting,” Journal of Japanese Society for Artificial Intelligence, Vol. 14, No. 5, pp: 771-780, 1999.
[12] J. Zhu, S. Rosset, H. Zou, and T. Hastie, Multi-class adaboost, Technique
Report, 2005.
[13] C.-W. Chen, 多類別AdaBoost的穩健性研究, Department of Mathematics, National Chung Cheng University, 2006.
[14] C.-H. Han and K.-B. Sim, ”Real-time face detection using AdaBoot algorithm,” International Conference on Control, Automation and Systems, pp. 1892 - 1895, 2008.
[15] M. J. Jones and J. M. Rehg, “Statistical color models with applications to skin detection,” IEEE International Conference on Computer Vision and Pattern Recognition, pp. 274-280, 1999.
[16] R. L. Fante, ”Central limit theorem: use with caution,” IEEE International Conference on Aerospace and Electronic Systems, pp.739-740, 2001.
[17] K. Kim and G. Shevlyakov, ” Why Gaussianity?” IEEE International Conference on Signal Processing Magazine, pp. 102-113, 2008.
[18] Y. Zhang, H. Lu ,and C. Xu, ”Collaborate ball and player trajectory extraction in broadcast soccer video,” International Conference on Pattern Recognition, Paper TuAT 10.23, 2008.
[19] G. Zhu1, C. Xu2, Y. Zhang3, Q. Huang4, and H. Lu3, ”Event tactic analysis based on player and ball trajectory in broadcast video,” international conference on Multimedia, pp. 58-67, 2007.
[20] P.-N. Tan1, M. Steinbach2, and V. Kumar3, Introduction to Data Mining, books, Michigan State University1, University of Minnesota2, University of Minnesota3, 2005.

指導教授

唐之瑋(Chih-Wei Tang)

審核日期

2009-7-20

推文