由於資料的普及、多媒體的盛行和網路的發達，我們可以經由許多的管道接觸到許多的多媒體文件，而最常見的多媒體文件就是影像。這些影像在不同的應用上可以形成許多不同種類的多媒體系統。當我們在使用這些多媒體系統時，其中的步驟不外乎是尋找和比對。但是尋找和比對影像是無法用單一的方法就可以解決的，而且相近似影像的判斷也會因人而異。我們不使用傳統的方法，即對影像先取出許多的特徵值，然後在特徵空間上的距離決定相似的檔案。我們把影像以較大的區塊為單位，把每一個新單位以新的數值來代表他，形成另一種特徵值。因為影像的形成本來就有其空間上的關係，所以我們都以空間的概念來形成我們的方法。我們利用較快速的投影方法，先找出影像區塊可能所在的位置，並利用原來影像裡的空間關係來去除不相似的檔案。再來我們利用動態時間偏移(Dynamic Time Warping)的概念導出動態二維偏移(Dynamic 2-Dimension Warping)的方法，利用較準確的動態二維偏移方法找出影像裡相似區塊的位置，並以和相似區塊的距離，作為最後輸出最相似影像的依據。我們在校園導覽系統中實作我們的方法。以實驗結果來看，動態二維偏移可以找到最相似的區塊，而且我們可以有九成的準確度找出最相近的地點。我們所導出的動態二維偏移方法，可以再延伸成為動態三維偏移方法或是動態多維偏移方法。

We can get more and more multimedia data in recent year. The major part of multimedia data is image. We can use image to build many type multimedia systems. Images are formed by many objects which have relations on space. We use Projection method to find where the blocks are on other images, and use the relations on space to filter out the image which can’t be matched. We use the concept of Dynamic Time Warping method to evolve into our Dynamic 2-Dimension Warping method. The method can precisely find where the blocks are. And we use the distance of blocks to determine the most matched image. The Dynamic 2-Dimension Warping method can be evolve into Dynamic 3-Dimension Warping method or Dynamic n-Dimension Warping method.

[1] Antonios Deligiannakis, N.R., Extended Wavelets for Multiple Measures. in SIGMOD, (San Diego, CA, 2003).
[2] Bin Cui, B.C.O., Jianwen Su, Kian-Lee Tan, Contorting High Dimensional Data for Efficient Main Memory KNN Processing. in SIGMOD, (San Diego, CA, 2003).
[3] C. H. Wang, H.C.L., C. C. Shih, H. R. Tyan, C. F. Lin Querying Image Database by Video Content. Journal of Information Science and Engineering, 19. 967-987.
[4] Chee-Yong Chan, M.G., Rajeev Rastogi RE-tree: an efficient index structure for regular expressions. VLDB, 12. 102-119.
[5] Cheng-Chia Chang, J.-S.R.J., and Ling-Ling Wang, Color Image Retrieval Based on 2D Strings. in Computer Vision, Graphics, and Image Processing, (Taiwan, 1998).
[6] Deok-Hwan Kim, C.-W.C., Qcluster: Relevance Feedback Using Adaptive Clustering for Content-Based Image Retrieval. in SIGMOD, (San Diego, CA, 2003).
[7] Ertem Tuncel, H.F., Kenneth Rose, VQ-Index: An Index Structure for Similarity Searching in Multimedia Databases. in ACM Multimedia, (2002).
[8] Feng Jing, M.L., Hong-Jiang Zhang, Bo Zhang, An Effective Region-Based Image Retrieval Framework. in.
[9] Hans-Peter Kriegel, S.B., Peer Kroger, Martin Pfeifle, Matthias Schubert, Using Sets of Feature Vectors for Similarity Search on Voxelized CAD Objects. in SIGMOD, (San Diego, CA, 2003).
[10] Jun Zhang, M.Z., Dimitris Papadias, Yufei Tao, Dik Lun Lee, Location-based Spatial Queries. in ACM SIGMOD, (San Diego, California, USA, 2003).
[11] Keogh, E., Exact Indexing of Dynamic Time Warping. in VLDB, (Hong Kong, China, 2002).
[12] King-Shy Goh, B.L., Edward Chang, DynDex: A Dynamic and Non-metric Space Indexer. in ACM Multimedia, (2002).
[13] Lei Zhu, A.R., Aidong Zhang, Theory of Keyblock-Based Image Retrieval. in ACM, (2002).
[14] Liyuan Li, W.H., Irene Y. H. Gu, Qi Tian, Foreground Object Detection from Videos Containing Complex Background. in ACM Multimedia, (2003).
[15] Lok-Lam Cheng, D.W.C., Siu-Ming Yiu, Approximate String Matching in DNA Sequences. in Database Systems for Advanced Applications, (2003).
[16] Michail Vlachos, M.H., Dimitrios Gunopulos, Eamonn Keogh, Indexing Multi-Dimensional Time-Series with Support for multiple Distance Measures. in ACM SIGKDD, (2003).
[17] Ronald Fagin, R.K., D. Sivakumar, Efficient similarity search and classification via rank aggregation. in SIGMOD, (San Diego, CA, 2003).
[18] Seok-Lyong Lee, S.-J.C., Deok-Hwan Kim, Ju-Hong Lee, Chin-Wan Chung, Similarity Search for Multidimensional Data Sequences. in IEEE Data Engineering, (2000).
[19] Simon Sheu, J.S., Effective Filtering for Nearest-Neighbors Queries in Large Time-Series Databases. in.
[20] Simon Tong, E.C. Support Vector Machine Active Learning for Image Retrieval. ACM Multimedia.
[21] Tamer Kahveci, A.K.S., An Efficient Index Structure for String Databases. in VLDB, (Roma, Italy, 2001).
[22] Tamer Kahveci, A.K.S., Aliekber Gurel, Similarity Searching for Multi-attribute Sequences. in.
[23] Xiang Sean Zhou, I.C., Qi Tian, Thomas S. Huang, Feature Extraction and Selection for Image Retrieval. in.
[24] Xiang Sean Zhou, T.S., Huang, Comparing Discriminating Transformations and SVM for Learning during Multimedia Retrieval. in ACM Multimedia, (2001).
[25] Xiang Sean Zhou, Y.R., Thomas S. Huang, Water-Filling: A Novel Way for Image Structural Feature Extraction. in.
[26] Yong-Sheng Chen, Y.-P.H., Chiou-Shann Fuh, Winner-Update Algorithm for Nearest Neighbor Search. in Pattern Recognition, (Barcelona, Spain, 2000).