隨著電腦技術之進步，影音壓縮格式在日常生活中普遍可見，多媒體影音資料庫管理方式也越來越重要，而一般傳統文字之管理方式不適用於影音資料管理，有效影片資料庫必需具備影片摘要，影片摘要內包含許多關鍵影像，關鍵影像是一種簡單又有效之方式代表一段影片之內容摘要，數張關鍵影像能完全表現出這段影片所要表達之內容，所以影片摘要在大量之影片資料庫下可幫助使用者快速瞭解影片內容並且有效地找出感興趣之影片內容。在某些情況下，既定時間限制下、儲存空間限制下或網路頻寬限制下影片觀賞，影片摘要以不同比率呈現，關鍵影像之數量在限制之下並且擷取最具代表性之關鍵影像，結合以上因素，影片摘要對於多媒體管理是一項重要之議題。
在影片摘要中，關鍵影像數目與影片摘要和原影片序列之間之失真度有關，影片摘要比率越高，影片摘要和原影片序列之間之失真度越小；反之，影片摘要比率越低，失真度越大，在本篇論文中，著重在以比率-失真度準則找出最具代表性之影片摘要，在不同影片摘要比率，取出與原影片序列之間最小失真之關鍵影像，每張關鍵影像都代表一小段影片，展示整部影片之內容結構，使用Normalized graph cuts(NCuts)分群法將相似影片段落分在同一群，分群之結果與時間資訊形成一個有向時間圖(directed temporal graph)，在有向時間圖上，最短路徑演算法找出整部影片之主要故事架構，最後實驗部份，Open Video Project所蒐集之影片作為測試影片，本論文影片摘要方法與Open Video Project所提供之關鍵影像以及以PME為主之方法做一個有意義之比較。

Due to advanced in computer technology，video data are becoming available in the daily life. The method of managing Multi-media video database is more and more important，and traditional database management for text documents is not suitable for video database; therefore, efficient video database must equip video summary. Video summarization contains a number of key-frame and the key-frame is a simple yet effective form of summarizing a video sequence and the video summarization help user browses rapidly and effectively find out video that the user wants to find. Video summarization except extraction of key-frame has another important key, the number of key-frame. When storage and network bandwidth are limited, the number of key-frame must conform to the limit condition and as far as possible find the representative key-frame. Video summarization is important topic for managing Multi-media video.
The number of key-frame in video summarization is related to distortion between video summarization and original video sequence. The number of key-frame is more, the distortion between video summarization and original video sequence is smaller. This paper emphasizes key-frame extraction and the rate of key-frame. First the user inputs the number of key-frame and then extracts the key-frame that has smallest distortion between original video sequence in key-frame number limit situation. In order to understand the entire video structure，the Normalized the graph cuts(NCuts) group method is carried out to cluster similar video paragraph. The resulting clusters form a direction temporal graph and a shortest path algorithm is proposed to find main structure of video. The performance of the proposed method is demonstrated by experiments on a collection of videos from Open Vide Project. We provided a meaningful comparison between results of the proposed summarization with Open Vide storyboard and the PME based approach.

目錄
第1章 簡介 9
第1節 序論 9
第2節 相關研究 14
1-2-1 分鏡變換偵測 14
1-2-2 分鏡合併 20
1-2-3 擷取關鍵影像 28
第2章 理論基礎 34
第1節 HSV模型 34
第2節 比率與失真度 36
第3節 分群法 39
2-3-1 Normalized Cuts 分群演算法 39
2-3-2 最佳化分割 41
第3章 研究方法 46
第1節 基以比率失真度準則之影片摘要 47
3-1-1 比率失真度準則 47
3-1-2 初始分割 49
3-1-3 擷取關鍵影像 50
3-1-4 NCuts分群法 55
3-1-5 時間情境圖 56
第4章 實驗結果 59
第1節 測試序列與實驗方法 59
4-1-1 測試影片序列 59
4-1-2 摘要評估 59
4-1-3 與其他演算法比較 61
第2節 實驗結果與分析 63
第5章 結論與未來展望 71
第1節 結論 71
第2節 未來展望 71
參考文獻 72

[1] H.J. Zhang, A. Kankanhalli, and S.W. Smoliar, “Automatic Partitioning of Full-motion Vide,” ACM Multimedia Systems , vol. 1, no. 1, pp. 10-28, 1993.
[2] R. Zabih, J. Miller, and K. Mai, “A feature-based algorithm for detecting and classifying production effects,” ACM Multimedia Systems, vol. 7, no. 2, pp. 189-200, 1995.
[3] Z. Cernekova, I. Pitas, and C. Nikou, “Information theory-based shot cut/fade detection and video summarization,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 16, pp. 82-91, 2006.
[4] Z. Rasheed, and M. Shah, “Scene Detection In Hollywood Movies and TV shows,” In Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 2, pp.343-348, 2003.
[5] H.B. Kang, “A hierarchical approach to scene segmentation,” Content-Based Access of Image and Video Libraries, pp.65-71, 2001.
[6] T. Kanungo, D.M. Mount, N.S. Netanyahu, C.D. Piatko, R. Silverman and A.Y. Wu, “An Efficient K-Means Clustering Algorithm : Analysis and Implementation,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 24, no. 7, pp.881- 892, 2002
[7] A. Hanjalic, and H.J. Zhang, “An integrated scheme for automated video abstraction based on unsupervised cluster-validity analysis,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 9, no. 8, pp. 1280-1289, 1999.
[8] L.H. Chen, Y.C. Lai, and H.Y. Liao, “Video Scene Extraction Using Mosaic Technique,” In Proceedings of International Conference on Pattern Recognition, pp.723-726, 2006.
[9] M. Irani and P. Anandan, “Video indexing based on mosaic representations,” Proceedings of the IEEE, vol. 85, no. 5, pp. 905-921, 1998.
[10] Coding of moving pictures and associated audio for storage media up to 1.5 m/s: Video, ISO/IEC DIS 11172-2, 1991
[11] Information Technology Generic Coding of Moving Pictures and Associated Audio: Video, ISO/IEC 13818-2, 1994
[12] ISO/IEC JTC1/SC29/WG11 14496-2, “Amd X, Coding of Moving Pictures and Audio,” International Standard, Maui, HI, 1999
[13] H. Yi, D. Rajan, and L.T. Chia, “A motion-based scene tree for compressed video content management,” Image and Vision Computing , vol. 24, no. 2, pp. 131-142, 2006.
[14] T. Liu, H.J. Zhang, and F. Qi, “A novel video key-frame-extraction algorithm based on perceived motion energy model,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 13, no. 10, pp. 1006-1013, Oct. 2003.
[15] Y. Zhao, T. Wang, P. Wang, and Y. Du, “Scene segmentation and categorization using NCuts,” In Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. 1-7, 2007.
[16] J.B. Shi, and J. Malik, “Normalized cuts and image segmentation,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 22, no. 8, pp. 888-905, 2000.
[17] C.W. Ngo, Y.F. Ma and H.J. Zhang, “Video Summarization and Scene Detection by Graph Modeling,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 15, no. 2, pp. 296-305, 2005.
[18] M.M. Yeung and B.L. Yeo, “Time-constrained clustering for segmentation of video into story units,” In Proceedings of International Conference on Pattern Recognition, vol. 3, pp. 375-380, 1996.
[19] Y.F Ma, and H.J. Zhang, “A Model of Motion Attention for Video Skimming,” In Proceedings of International Conference on Image Process, vol.1, pp. 129-132, 2002.
[20] P. Mundur, Y. Rao, and Y. Yesha, “Keyframe-based Video Summarization using Delaunay Clustering,” International Journal on Digital Libraries, vol. 6, no. 2, pp. 219-232, 2006.
[21] Z. Li, G.M. Schuster, A.K. Katsaggelos, and B. Gandhi, “Rate-Distortion Optimal Video Summary Generation,” IEEE Transactions on Image Processing, vol. 14, no. 10, pp. 1550-1560, 2005.
[22] S.Y. Huang, C.Y. Cho, and J.S. Wang, “Adaptive Fast Block-Matching Algorithm by Switching Search Patterns for Sequences With Wide-Range Motion Content,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 15, no. 11, pp. 1373-1384, 2005.
[23] S. Zhu and K.K. Ma, “A New Diamond Search Algorithm for Fast Block Matching Motion Estimation,” IEEE Transactions on Image Processing, vol. 9, no. 2, pp. 287- 290, 2000.
[24] G.H. Golub, and C.F. VanLoan, Matrix Computations. John Hopkins University Press, 1989.
[25] The Open Video Project. http://www.open-video.org/