臺灣博碩士論文加值系統

在DV日漸盛行之下，以video作為應用的領域也隨之增長。以往研究從擷取影像shot，發展至abstraction、 summarization等剪輯技術用於表現影片中精采的片段資訊，甚至增加audio技術讓使用者能隨著音樂的節奏觀賞精采鏡頭。然而在眾多的發展中，仍以單台攝影機作為基礎，因此在拍攝影片時會因而產生表演內容與人物特寫無法同時被擷取的狀況，因此以多台攝影機作為拍攝的工具可以達到預期的目的，讓使用者能拍攝更豐富的內容與特寫。
本論文研究是以多視角影片為基礎，建立自動化的視訊編輯系統，予以解決多台攝影機拍攝之影像重合的問題。多攝影機視訊合成的問題有(1)多視角影片的同步化( video synchronization )，(2)對於影片的轉換( video switching )作為主要的討論主軸。多視角影片的同步化是把影片中本身的時間對應到共同的時間軸，在結合成影片的同時不會發生重疊的問題。本系統首先利用Abrupt video shot detection切割影片中發生瞬時變化的片段，並利用於我們所設計的video synchronization技術讓系統能搜尋到同步時間點。因為在不同角度下拍攝的人，所表現出來的速度會有相似的狀況，因此在此技術中利用velocity curve作為偵測上的特徵，並利用相似度的比較達到正確的偵測判斷。轉換影片的目的是以擷取不同影片內容讓使用者能提高興趣並能觀看到當時的重點鏡頭。本系統設計三種以內容為基礎的shot，考慮使用者對於影片所關注的部份而分成camera motion shot、face shot與fragment shot，並利用我們設計的video switching計算每段shot的重要性予以判斷是否會被結合於最終的影片。
本研究實驗以球類比賽為拍攝內容。對於拍攝角度上的問題、content-based shot權重的不同、室內室外環境的改變與拍攝多人影片的問題做討論，並分析影片的同步化與shot在不同狀況下的重要性。本論文提出方法以解決多影像編輯所發生的同步化與影片轉換問題，並設計利用velocity curve、content-based shot與importance shot value來完成更順暢且活潑的影片。

As digital videos become more and more popular nowadays, the application of videos in different fields has been wide spreading. The researches had turned from simply getting the shots to using techniques like abstraction, summarization to display the exciting performances in the clips; or even increasing audio techniques to let users listen to music while watching the films. However, among these developments, the filming still relies on only one video camera. When filming, it makes it impossible to catch the performance and a close-up on the performers at the same time. Therefore, using multiple video cameras to film can achieve the expectations of letting the users capture bounteous contents and close-ups.
The thesis takes multi-view based video as the foundation to build up the automatic video editing system. There are two important issues for video composition: video synchronization and video switching. Video synchronization is to match the time of the videos from different viewing directions to the global time axis. The system firstly uses abrupt video shot detection to segment the abrupt shots of the captured videos, and then uses the velocity curve similarity to search the synchro-point. The goal of video switching is to retrieve different contents of videos to appeal the users and allow them to watch the attention shots of videos. We designed three shots that based on the contents for the system, and we categorized these three shots in considering the parts that users will take notice of on videos, for example: camera motion shot, face shot, and fragment shot; we calculate the importance of each shot to determine whether those shots should be selected into the compositive film or not.
The experiments uses ball games as the filming content, and they contain the conditions of the different viewing angles, different content based shot weighting, the environmental change of indoors and outdoors, and the filming of a lot of people. Also, we analyze the synchronization of the film and the importance of the shots in different circumstances.

摘要.............................................ii
Abstract........................................iii
Content..........................................iv
List of Figures..................................vi
List of Table..................................viii
Chapter 1 Introduction...........................1
1.1 Related Works............................1
1.2 System Framework.........................3
1.3 Thesis Overview..........................4
Chapter 2 Video Synchronization..................5
2.1 Abrupt Video Shot Detection..............5
2.1.1 Pixel-based Measurement..................6
2.1.2 Histogram-based Measurement..............7
2.1.3 Edge-based Measurement...................7
2.2 Video Synchronization Flow Chart........10
2.3 Background Construction.................11
2.4 Human Tracking..........................12
2.4.1 Background Subtraction Method...........13
2.4.2 Human Tracking in the Primary Video.....14
2.4.3 Human Correspondence....................15
2.5 Human Velocity Construction and Synchro-Point Estimation.......................................17
Chapter 3 Video Switching.......................20
3.1 Content-based Video Shot Detection......20
3.1.1 Camera Motion Detection.................21
3.1.2 Face Shot Detection.....................23
3.2 Video Switching.........................26
3.2.1 Intra-Importance Shot Comparison........27
3.2.2 Inter-Importance Shot Comparison........30
Chapter 4 Experiments...........................32
4.1 Environment Setting.....................32
4.2 Experiments and Discussions.............33
Chapter 5 Conclusions...........................51
References.......................................52

[1] Chung-Lin Huang and Bing-Yao Liao, “A Robust Scene-Change Detection Method for Video Segmentation”, IEEE Trans. on Circuits and Systems for Video Tech., Vol. 11, No. 12, Dec. 2001
[2] Hong Lu and Yap-Peng Tan, “An Effective Post-Refinement Method for Shot Boundary Detection ”, IEEE Trans. on Circuits and Systems for Video Tech., Vol. 15, No. 11, Nov. 2005
[3] Adriana Dumitras, and Barry G. Haskell, “A Look-Ahead method for Pan and Zoom detection in Video Sequences using Block-Based Motion Vectors in polar coordinates”, In Proceedings of the International Symposium on Circuits and System, Vol. 3, 23-26 May 2004
[4] Douglas Chai and Abdesselam Bouzerdoum, “A Bayesian Approach to Skin Color Classification in YCbCr Color Space”, In Proceedings TENCON, Vol. 2, 24-27 Sept. 2000
[5] Son Lam Phung, Abdesselam Bouzerdoum, and Douglas Chai, “Skin segmentation using Color and Edge Information”, In Proceedings. Seventh International Symposium on Signal Processing and Its Applications, Vol. 1, 1-4 July 2003
[6] Vladimir Vezhnevets, Vassili Sazonov, Alla Andreeva, ”A Survey on Pixel-Based Skin Color Detection Techniques”, Proc. Graphicon-2003, pp. 85-92, Moscow, Russia, Sep. 2003.
[7] Hedvig Sidenbladh, ”Detecting Human Motion with Support Vector Machines”, In Proceedings of the 17th International Conference on Pattern Recognition, Vol. 2, 23-26 Aug. 2004
[8] Jin Young Kim, ”Fast Multi-face Detection in Color Images using Fuzzy Logic”, In Proceedings of International Symposium on Intelligent Signal Processing and Communication Systems, 13-16 Dec. 2005
[9] Min Xu. Ling-Yu Duan, Chang-ShengXu, Qi Tian, ”A Fusion Scheme of Visual and Auditory Modalities for Event Detection in Sports Video”, In Proceedings of the IEEE International Conference an Acoustics, Speech, and Signal Processing, Vol. 1, 6-9 July 2003
[10] Ling-Yu Duan, Min Xu, Qi Tian Chang-Sheng Xu Jesse S. Jin, ”A Unified Framework for Semantic Shot Classification in Sports Video”, In IEEE Trans. on Multimedia, Vol. 7, No. 6, Dec. 2005
[11] Hong Lo aud Yap-Fcng Tan, “Content-based Sports Video Analysis and Modeling”, Seventh International Conference on Control, Automation, Robotics and Vision, Vol. 3, 2-5 Dec. 2002
[12] Wen-Nung Lie, Ting-Chih Lin, and Sheng-Hsiung Hsia, “Motion-Based Event Detection and Semantic Classification for Baseball Sport Videos”, In IEEE International Conference on Multimedia and Expo, Vol. 3, 27-30 June 2004
[13] Ahmet Ekin' and A. Murut Tekalp', “Shot type Classification by Dominant Color for Sports Video Segmentation and Summarization”, In Proceeding, IEEE International Conference on Acoustics, Speech, and Signal Proceeding, Vol. 3, 6-10 April 2003
[14] Hong Lu and Yap-Peng Tan, ”Sports Video Analysis and Structuring”, In IEEE Fourth Workshop on Multimedia Signal Proceeding, 3-5 Oct. 2001
[15] Navneet Dalal and Radu Horaud, “Indexing Key Positions between Multiple Videos”, In IEEE Proceedings of the workshop on Motion and Video Computing, 5-6 Dec. 2002
[16] Surk-Woo Jang, Gye-Young Kim, Hyung-ll Choi, ”Extracting Velocity Information of Multiple Moving Objects”, In Proceedings of the IEEE Intelligent Vehicles Symposium, 19-20 Sept. 1996
[17] Jianpeng Zhou and Jack Hoang, ”Real Time Robust Human Detection and Tracking System”, In Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Vol. 3, 20-26 June 2005
[18] Cavan Reilly,, Phillip Price, Andrew Gelman, and Scott A. Sandgathe, “Using Image and Curve Registration for Measuring the Goodness of Fit of Spatial and Temporal Predictions”, In Biometrics 60, 954–964 Dec. 2004
[19] Hemant D. Tagare, ”Deformable 2-D Template Matching Using Orthogonal Curves”, In IEEE Transactions on Medical Imaging, Vol. 16, Issue 1, Feb. 1997
[20] Tang Hong, Fang Tao, Shi PengFei, “Nonlinear Spectral Similarity Measure”, In Proceedings of the IEEE Geoscience and Remote Sensing Symposium, Vol. 5, 2004
[21] Chen-Hsiu Huang, Chi-Hao Wu, Jin-Hau Kuo, Ja-Ling Wu, “A Musical-driven Video Summarization System Using Content-aware Mechanisms”, In IEEE International Symposium on Circuits and Systems, 23-26 May 2005
[22] Ashfaqur Rahman, Manzur Murshed, and Laurence S. Dooley, ”Feature Weighting Methods for Abstract Features Applicable to Motion based Video Indexing”, In IEEE Proceedings of the International Conference on Information Technology: Coding and Computing, Vol. 1, 2004